Transcripción automática, sin corregir a mano. Puede tener errores.

00:00 all right ladies and gentlemen I think we still have a couple of open seats at the bottom there for those looking for seats but without further Ado we get started um good morning well rather good afternoon I'm Daniel Yol Program Coordinator at Berkman Client Center um just wanted to relay a couple announcements to youall about our some of our upcoming events on Thursday we have litigating free speech cases in the African Regional courts featuring Brickman Klein fellow Nanny Jensen rento so that will be in Hower and our next lunch and talk will not be next Tuesday but the following Tuesday November 15th and that's the end of ownership with Professor um Aaron panowski of Case Western uh Reserve law school so without further Ado I before I introduce our introducer of our speaker I just want to remind you all that this production is live webcast for Prosperity it'll be available um on

01:00 YouTube within the next couple of days so uh be mindful of what you say and yeah that's or not that that's ad option as well so now just great thanks everybody thanks so much for being here I'm going to be really brief and then hand it off to pal just by saying that I run the clinical program here at the Berkman Klein Center we work a lot with the ACLU of Massachusetts and paa first came her work first came to my attention through that connection to ACLU the data for justice project that she's going to be talking about about today is an outgrowth of the work that she's been doing at the ACLU the ACLU of Massachusetts over the last year as a Ford Mozilla open web fellow um and I've seen her write about her development of a data visualization framework I think we're going to hear a lot about that and about more broadly about the themes of data science in the public interest and without further Ado I'll just hand it off to thanks and to be clear we we'll save plenty of time for questions at the end we have a couple microphones so we can make sure everything is captured um so just at the and we'll we'll hand

02:01 those around and uh do some Q&A thanks thank you uh can you hear me okay yeah cool so thanks all for coming uh I'm honored uh to be here talking to you uh especially in this amaz in this topic that is very interesting in my opinion let me just put my password okay um presentation okay so the topic today is public interest in the uh data science and I want to start by saying that uh well with this quote which is able lawyers have to to a large extent allow themselves to become agents of great corporations and have neglected their obligation to use their powers for the protection of the people um this quote started the public interest law and I think if if we change lawyers and we use technologist the same

03:04 applies so this is what I'm going to talk about today um and but first I want to Define what is public interest and I think it's very simple it's to affect change in public policies uh in the interest of the Commonwealth um and to advocate for the under represented and build this or or cheer for the underdog so to speak and then what is data science data science is um to extract knowledge um and insight from information and data and then it's the intersection of math statistics research information science and computer science and if we take that into consideration it might be very simple to explain uh what um what what the public interest data science is so what happens like the the motivation of why I I applied for these fellowships and why I'm working on what I'm working on is because I was wondering what uh

04:06 will happen if you put data and technology in the hands of uh civil liberties organizations human rights activists uh media outlets and governments so my background is a little bit uh in the government in the open data movement and the open source software movement so I I have lived um like my experience is uh just in a few words is that I wouldn't be here if it were not for open source and the internet and the open web so I think that data and Technology as it changed my life it can also change society and I think we are going to uh see that more in the future um so now that we recognize that there are civil liberties organizations or human rights activists that need data and Technology um but why what what can you can we change um oh sorry they want me to move here yeah

05:09 okay let's I need to control this okay yeah this is better right yeah I'm fine but I have a mic here hey hello hello hello hello yep cool okay um um so what do what do they need data for and I think it's oh it's okay uh we can affect change in Liberty and Justice and equality and especially transparency and accountability and I think these are the main things that we can um effect change in and I think that's the key for uh like a new Society so but but who else can data help and I think lawyers journalists policy makers and Advocates and activists can benefit from data and

06:10 Technology but um maybe this is like a very uh hard uh like they need the skills uh they need to gather the skills for them to have an effect on their causes and that's basically why I think this is very important so the I think also that intersection of data and liberty for instance or data and Justice and data and and equality can have U very important impact in in privacy or surveillance or or ending the war on drugs or the mass incarceration um phenomenon and then we can also through data and Technology uh and um or at least shed light in biases and improve uh access to justice for uh a lot of people especially the under the under represented people so I'm going to present the data for justice project which is a

07:10 collaboration well it's enabled by the Ford Foundation the ACLU of Massachusetts and mosil Foundation um it started as my as the project of my fellowship like the open web Fellowship uh that I was in 2015 and it became now um like my fellowship got extended at the ACLU and now as as a a Berman Center fellow um I'm trying to uh absorb uh bergman's knowledge into this project so the data for justice project aims to empower Empower lawyers and Advocates um to make fact-based and data driven decisions and uh do their work like advocacy um to improve access to Justice and their communities by providing them and decision makers with answers to questions such as how isolation bias policing uh race and poverty relate to each other uh what we're seeing here uh

08:13 and I'm going to talk about that um more in depth in the in the following minutes is this is Boston Massachusetts and what you're seeing is the green color is people of color uh obviously The Greener the color uh the more percentage of the higher the percentage of people of color live in that area and the red dots are um uh drug arrests so this is like only a a simple example of how data and data visualization in this case can show that basically um the water on drugs is racially charged so I'm going to talk about the process um and how we try to implement this process at the ACLU and I think the first thing and it's the most important is how do how do you get the data um we

09:14 have here an Lambert which is uh uh one of the lawyers at the ACLU and she is very uh like this is her topic like how do you get data and I think uh Freedom of Information Act um are the key but in many cases especially coming from the ACLU they get denied like so in one of these cases that I'm going to talk about um we ask for police records and we we we waited for a year and a half and they will give it to us so we so and also it's very important to uh don't give up on getting the data because you want to change uh this uh problem so we stood and we won um and then uh after that I normalize the data and that's that's my work like the the the things involved are my work I normalize it and process it and and analyze it and visualize it and but that's not it it doesn't end

10:15 there I think we also have oh we have Kate here which is an advocate at the ACLU and their work is um to socialize it and inform change and then we also have other departments that uh are uh that need their work is to make it last so if we use it as a flowchart it's okay you identify a cost uh and if you have data well J data uh if you don't have data which is very common um do you know who who has data if you don't know well you need to research and if you do know who has it and it's is it is it open data um if it's not well you need to go through the legal channels which which are the ones I metion mentioned foas and lawsuits and then you get the data and after you get the data uh this is my work this is data science uh you do you

11:18 develop um measurements and indicators and you research and measure and you asked your question is this good is this good is this good is until it's not and when it's not you can design a solution or a or a methodology to solve or improve this problem with Community engagement and this is where Kate's work uh starts and yeah and and I think it's very important also from the beginning to involve communities in the solution so it's more like a um co-designed uh solution because it's very easy for us uh uh that have the data and the skills to say oh I'm able to solve the problem by myself and I'm uh going to reach out to communities with the solution and uh I think that's not how it should work especially when you are talking about people um affected communities should be involved from the beginning of the and

12:20 that's what the ACLU basically does but then um if you want to change these things you need to also involve all the actors and stakeholders so you also need to codesign with them and you need to uh create uh run workshops and do uh this co-designing for the solution uh I think it's very important that the government for instance is it's also involved which is the hard part because in in many ways you are saying to the government or to the public officials that their work is not done as good as it could be right and especially in the case of the criminal justice um system so um you have the your workshops and do you have a solution do you have an idea that can be implemented do you have like practical uh or empirical um solution that can be implemented and

13:21 if you do well yeah you you can relax a little bit and then make it last make it uh through legislative means or to other means to make it permanent I think it's also very important to think about that but it's like a separate Pro process and it takes longer and longer and longer so this is how it looks like all the this is the the I call it the data process and it's like an abbreviation and this how I I look the process of the data for justice project um so I'm going to speak about the projects uh like the particular projects we we've been working on uh here so we have three projects right now which can be accessed the first one uh is the Duan case how many of you are familiar with Duhan okay oh cool so I'm I'm going to go inside in deep on that but we also have stop and frisk data sets and the

14:22 marijuana reg uh regulation uh that is coming in the ballot in November 8th um so as you can see it's basically work between the legal department and the advocacy Department um and data is in the intersection of of both so the Duan case um an Duan was a chemist that worked in a in a crime lab in Massachusetts and she was um in charge of Performing the drug tests for Contraband uh or seizures uh she falsified um evidence and thanks to data and thanks to the work that the ACU we have come to the number of more than 24,000 cases so these are 24,000 people that didn't have access to justice that uh the that their rights were violated because the um the I I forgot the the word um okay

15:28 um so yeah their rights were violated because of this uh action of the chemist she was working at least for six years like six complete years and then um two more years that were half and half so the the impact is that a lot of people were Inc incarcerated because of her um and some some say that there that that there's no more people incarcerated because of her but still the impact is that some people were deported others have bad records that can also affect in other jurisdictions so for instance there's a case in Florida where a man is facing life in prison because um he committed a crime or in in Florida and the prosecutors um find out one of these cases and that will and that is the his uh third strike so that that can happen and also um one of the key key of the

16:30 keyy uh things to know about this case is that prosecutors say that these are bad bad people that they were like floting the the streets with drugs and in fact only 62% uh I mean in fact 62% of the cases are possession only so it's uh consumers instead of Distributors uh that are affected by this um so only 30 37% that's distribution so we have uh some interesting um tables so this table uh talks about um is a division of the 24,48 4483 cases divided by by County uh and also divided by adult cases and jinile jinile cases um as you can see uh Essex and sok are the top two uh clients of this this uh this lab and

17:35 then uh this is the the distribution of uh the type of cases so 62% are possession and distribu uh no distribution and 37% are other are distribution and 1% is other cases which are like uh conspiracy and others that were not POS possession or distribution and then um we have this table that shows uh the percentage of cases with adverse disposition um from 2013 and the percentage and what percentage they represent in the total of the cases so we have some counties that are that almost 30% of their drug cases were affected by this for instance um it's uh I think it's nuket or what is this not norol norol U has 30% of their cases tainted and suul has 32% of its cases um

18:37 tainted that's a lot of people um so yeah um you can access the this Aid this are screenshots from the affidavits I submitted uh that we that I worked with the lawyers at the ACLU um so I I have the link uh that I can send it to you because they are like very interesting and it's this is only like a light resume of this case so yeah I think that we can actually use data to change these 24,000 people lives and I think it's very powerful um then another example of this work is and then I we can discuss and all that uh it's the yes on four which is the question four um and if you say yes to the to if you answer yes on the question four in the November 8th ballot you are validating

19:38 the mar marijuana legalization well regulation and Taxation in Massachusetts um and the argument we are using and we are using data to prove it is that marijuana related police interactions um are racially biased and because they only happen mainly where people of color live um and you can see that it's um it's again the same the same pattern green color is people of color and the red dots are where arrest happened since August 2015 to date so yeah there's a I think a very clear pattern and I want to show you how this looks so I'm going to go go sorry I'm going to come here

20:50 okay oh where is it yay okay so this is an advocacy campaign that can be accessed by uh this website um which is yay yes okay so let me just play marijuana policing destroys lives every week in Boston even though Massachusetts overwhelmingly passed marijuana de Criminal ization in 2008 right now an average of 13 people every week are arrested or harassed by the police because of marijuana just one arrest can permanently change one's life for the worse causing barriers to employment schooling and housing the data shows the majority of police interactions and

21:51 arrests for marijuana offences occur in neighborhoods where mostly people of color live let's explore the data from the 2014 census and the data produced daily by the Boston Police Department 46% of Boston's population is white 25% of Boston's population is black and 29% of Boston's population are other people of color as you can see the city is racially segregated white people of color white people of color now let's look at where police harass and AR arrest people because of marijuana this is where marijuana incidents happened last week and this is where marijuana incidents happened since August 2015 do you notice a pattern white people of color we can help end

22:51 racially biased policing of marijuana vote Yes on four so I think that's uh um this is the one of the examples of what data um technology can help uh like to do this type of arguments so let's see how does this work and I'm going to talk about my tools um I can't find the mouse okay I think it's here no nope here it is oh no oh my God what is it everyone come okay here it is okay oh did you get the URL it can be accessed by your phone and all that so um share

23:52 it so this is my work these are my tools I do a lot of data cleaning and normalization and I think a lot about data structures um I love like uh I consider myself like a uh systems architecture architect and also a data architect more than a a data scientist I do a lot of structure thinking and I do a lot of processing I use tons of uh Python and PHP and SQL uh I've been working on SQL for almost I don't know uh 17 years so I I know it very well and that enables me to uh leverage these uh skills to uh actually make it very easy for us to to Output this processed and um process data to other tools like stata or R which are also the things I I use and then also I think one of the

24:54 main skills is uh GIS which is the the things I use for uh and I think they're crucial for um translating all these data to where people live or where people is so that's why I love posis I think it's very important uh tool and then to visualize it to to tell a story I think um is also very important and to do that I developed this framework which I'm not going to talk about that much but I developed a data visualization framework that enable Ables me to do that kind of pieces uh like the one of marijuana legalization uh in very little time and with a lot of information uh being displayed and analyzed so it's name it's uh ant which is which stands for augmented narrative toolkit and it it it's built on already developed tools like D3 and scroll magic and jQuery and

25:58 all these other tools I don't want to rain bend the wheel I just want to imple to make it easier for people to um to use these very complex tools like the three so my aim by developing andt is um for you to be able sorry for you to be able to put a map or a chart in a web page as easily as you would an image um which is like an standard HTML uh tag so my theory behind that is that if you enable people uh to use uh HTML uh to do data visualization HTML and CSS I think that we are going to start to see a datadriven um internet and I think that's that can be very very cool but then what uh the visualizations are cool um but I think the most important um

26:59 thing it's the Outreach we want to have impact on people's lives uh I mean the technical challenge is there and it's kind of cool to solve it but I think uh that if the work I've been doing doesn't have an impact on people's lives I think I wouldn't be happy about it like I will be only half happy so I think Outreach to communities Outreach to the authorities or the lawmakers it's very important too and so for instance there was a case where um one of these visualizations got the attention of the Urbano project which is um a youth of color um work group in Jamaica plane and what they asked their their fellows is to create this um to to do something with the with this data and what the kids did I think was like a an amazing thing was a they

27:59 painted the map that I just showed you uh in a mural in Jamaica plane and they marched with that mural um through all these neighborhoods and I think uh it's very low Tech but I think that's like the that completes a cycle of data uh because that specific those data points those specific data points went from um the asking for the data in the a Foya then suing for them for for it then it got Pro processed and we reached out to communities and the communities did something that then impact in in other types of uh stakeholders and I think when when I saw that I cried I was like oh this is this is how things should be working it's not about a technical challenge so my conclusions are that data and Tech do Don't Do Anything by themselves people do uh data

29:01 science is very cool very very cool and it's a technical challenge um but it's only a part of of of a team with a common goal I wouldn't be standing standing here if it were not for my co-workers co-workers at the ACLU that are kikass um and also the intersections are very important uh exploring law and data it's it's very cool exploring uh data and uh inequity is also very important so yeah data and Tech in the hands of the public interest organizations working to reduce inequality can have an impact on people's lives thank [Applause] you so if anyone has questions yay cool thanks

30:02 I have a question I love this presentation thank you so much so there was a point at the beginning of your of your talk okay thanks hi I was just saying that I love this project uh and all the work that you're doing uh with your colleagues and my question is about uh in the beginning of the presentation uh in your flowchart um you talked to there sort of a parenthetical statement where you said you could also involve the community that's most affected earlier in the process for example when uh in data Gathering or identifying the data sets that are needed and I know that Urbano project that you just referenced they did a a project like that where um they were looking at Transportation inequality in Boston and they had young uh youth of color from Boston keep journals of their transit times like to school to work that type of thing and they created visualizations of uh Transit inequality uh by race and geography in Boston so I wonder if you could talk a little bit more um about yeah what are what are the what are the

31:02 challenges that are keeping more data scientists or people who are doing the type of work that you are from engaging uh communities that are most impacted at different points in the process why does it happen so infrequently because I I think it's hard um it's it's expensive and um yeah I don't know if it were not for me being a in a paid Fellowship I will be able to do it uh because raising the money to compete with Facebook for the ACLU is a challenge and then um reaching out is it's a full-time job like it's it's not very easily done and you need to have like the street credibility and all that uh you need to have connections to the communities and I think uh that only through the collaboration with um with uh organiz ations you can do it like Community organizing for reals I think that's that's the main challenge

32:03 so data scientists are um very like technically skilled but they it's they're very far from in in I'm generalizing of course what uh from Community organizing great great project and the visualizations are compelling of course uh I wanted to so a lot of these data are are very personal and private there's data about race and uh criminal convictions and location and I'm wondering what's done in terms of uh privacy in the data sharing process do you get the raw data do you get like uh to just uh perform queries and what's the Privacy um cover for this well um so in the Duan case I had access to private information like personal information that that includes um names and Social

33:04 Security numbers of people uh like that were charged in these cases um this data is impounded and it's like I use security measurements to keep it that way and that's one case the other maps are open data and so they they've been through a process of the an anonymization and so for instance uh if I showed you this okay so these are not where exactly happened these are like a gri so a g that is I think uh half a mile uh so if an an incident happened there it goes It goes to that point that enables us to to remove the ident ification layer so people or for instance uh um so so

34:05 people don't can't say where something in particular happened so that they don't have the repercussions of being anonymized so yeah um paa I my name is Kate crockford I work with pal at the ACLU I just wanted to actually toot Pal's horn a little bit more than you did um to tell you all the sign significance of having a data scientist on staff for the legal department in the midst of the Bridgeman litigation so the Bridgeman litigation is the drug lab case we got the disposition data through that litigation as Pala said some of it was impounded we filed a motion asking the court to um unseal the non personally identifiable information aspects of that data which they did um and then paa was able to perform this really important analysis which revealed things like the fact that 62% of these cases were possession only which nobody knew before that um so that was reported in the press as a major finding we that was like a major part of

35:05 our brief one of our briefs in the case and then you know that one in three cases in suffk County was tainted by this Scandal I mean these are these are really important pieces of our legal arguments in um the Bridgman litigation that we would not have been able to make had we not had paa on our staff so I just want that to be very clear and also I would just like to tell you I don't know if you know this but the prosecutors filed a motion couple last week that pretty much disses paa really hard and says which I think is great because it means you're doing the right thing people powerful people hate you already so fantastic um they said they said things they effectively said I mean if you want the Bridgman litigation is really insane I encourage you to read about it because it'll put your head in a spin cycle so one of the things they said about paa is that it was not the intention of the Court when they unsealed this information for anyone to actually look at it they really said that the prosecutor said that in a brief

36:05 to the court yes the court unsealed this information but the purpose of unsealing it wasn't so people could understand what was going on okay and we're really upset that this meddling pal of VI Royale person bothered to look at it so just so you know she's really making an impact it's really valuable from the ACLU to have a scientist working on information like this thank you yes hi is there a is there a page on the web anywhere where you uh link to all the tools that you use to do this uh so that other people can do the same thing and follow your methodology yeah we have a GitHub page I mean all this except for the Duan case which is like a court case and Ono litigation all other things are public in GitHub um um yeah um so GitHub come uh data for

37:07 justice dat Force yeah so you you you will be able to see the code for this which is in the this a GitHub page so you will be able to see this and the methodology that I need to update and this is like this describes the the process to get the from the Census Data to the police data and all that so and there are all PRS so yes we know what happened to the uh person who falsified the data but the the question is what happened to the people were in jail uh some some are um maybe an can uh answer that question in particular I'm sorry to put you on the it's okay um I'll try um we

38:09 don't we don't know the actual uh I don't know the actual numbers the response of the response of the legal system was to recognize immediately that the people who were serving time at that moment when this was discovered had to be have an opportunity to challenge their convictions and get out and in um pretty short order the court system mobilized and had uh special magistrates appointed special hearings all of this by the way to follow up on Cade's point was challenged by the Commonwealth oh you can't vest those retired judges with judicial powers to admit people to bail or even to to have hearings where they make recommendations to a real Superior Court Judge about releasing people on bail but the the litigation that's going on now and and who knows I mean I think

39:10 probably the court system did an a good um a firstcut job in figuring out who was still incarcerated and doing something about it the problem uh was uh as Poli has described uh that then a certain amount of inertia lassitude set in what about the 24,000 other people what what are the remedies going to be for them um should we do something or as the Commonwealth seems to say you know it's just so formidable you know finding these people what you know gee it's so over overwhelming you know we just can't handle that so the point of the Bridgeman litigation bolstered by what paa has been able to demonstrate is you know we have to do

40:11 something about it the burden of these uh tainted convictions shouldn't fall on the people who got them it should fall on the people who went about Prosecuting them and then imposing them so that's the issue that it's going to be argued in the court next week I mean that's semi an answer I mean it divides the world of duken victims into people who were incarcerated where the system sort of responded but not fully uh and the uh people that the system uh at least with you know Commonwealth versus John Smith the Commonwealth was kind of willing to let those go where the people in the Bridgeman case uh or organization the public defenders have said uhuh uh-uh you got to pay attention to them too yep

41:13 [Laughter] yep uh I've been conducting or hosting a discussion in the context of a Harvard xmo called jury X which began disc discussion just this week of the issue of yes on question 4 and the most provocative forb hello yes the most provocative question has been the the way a way how to think about the fact that the leadership of the Commonwealth in the person of the governor the attorney general and the mayor of Boston have been absolutely opposed to the yes on 4 and the question posed that was provocative is is there a relationship between the uniformity of

42:13 the opposition of law enforcement emphasized in this morning's paper by the support of the prosecutors for the opposition uh police opposition to yes on for is there a relationship between cannabis reform and black lives matter and you've answered that question your your work is such a beautiful demonstration of it thank you that it is remarkable that black lives matter as a movement hasn't cottoned on to cannabis reform as a sister issue I mean look here we are we're not a black lives matter group and yet we can see it absolutely clearly it's a curiosity it's a but thank you very much for the work you've done it's extraordinary thank you

43:17 well thanks hi paa I've been reading um Kathy O'Neal's new book weapons of math destruction uh which is a terrific look at how algorithms are impacting uh vulnerable communities around the country oh sorry so it's a book called weapons of math destruction it's written by this woman Cathy O'Neal who was at De Shaw and as an implied math person um and I was thinking about that and about your brandise quote from the beginning and I'm wondering if you can talk about um the origins of your passion for social justice as a data scientist and what we can do in the data science community to get um data scientists excited about this do we need some sort of formalized training the way we do pro bono for law students okay so to answer that question I will have to tell you like my background which is I'm Mexican I'm 32 years old and when I was 15 I started working well I learned

44:19 to code when I was 12 and I only was able to do that because I I had access to the internet and I started talking to people that were coincidentally uh also developing the chats we were using so that um I don't know that picked my curiosity um so that's how I learned about Linux back in 1997 and then I learned the about C the the language so fast forward three years I started working as a web developer in Mexico when I was 15 and then fast forward work for years I was the C admin for Mexico's president's office and I was in charge of securing the web servers and enable enabling uh like a development platform for uh the Mexico's president's office um so I started reflecting on what had enabled me to do that because at that moment I recognized that I didn't have like peers my age uh

45:22 doing the work I was doing so the answer was open source and openness so I think that and that's when I decided that um through openness and open source uh I could actually enable people more people to to do the work I'm doing and now fast forward 12 years and I'm doing that I'm I've been working not in governments I've been working with um sexual and reproductive um uh organizations uh freedom of expression organizations um and all this type of work uh are only trying to demonstrate that openness and open source can actually change people's lives and now I think my my goal is to inspire Inspire other people to do just that so yeah I think that's that's it

46:30 I thank you for the um the ballot question on four now it does bring up a point of you know what in data analysis like this um there's the legal aspect that you dealt with and and certainly um you know that is very apparent now what is the what's the social impact on this type of data analysis where um cannabis reform is one part is the legal aspect of it the other part is the effect on the community as a whole what kind of data exists for gateway drugs um in in other communities where it has been decriminalized um that's what the social impact would be is there any kind of work along those lines because the the ballot question and I'm I'm speaking as a as a lone voter out here is flawed

47:33 because it it Deals Only with the criminalization but it is not it's it's actually about the long-term social impact that can happen in this community what has what is the data that exists elsewhere yeah do you want to answer that cave um sure I can try uh I think I hear you saying that if we tax and regulate marijuana more people will smoke marijuana is that what you're saying I'm not saying anything I'm just asking about the data itself that must exist not only in the United States but in in U neighboring countries where um decriminalization has taken place so that's what I'm asking about is that the data analysis has a social impact right I guess I'm asking what kind of data exactly data about use or data about data about use yeah and and whether whether in in terms of the whether cannabis is a drug that uh U

48:37 one shouldn't be bothered about you know as opposed to to heroin or crap or something like that well this is a little a field of the subject I think but just very quickly um there is data from Colorado that shows that um basically marijuana use rates don't go up significantly at all after legalization also we have passed decriminalization in Massachusetts so it's already not a crime to possess under an ounce of marijuana for personal consumption um and we haven't seen measurable increases in use since then what we have seen however is a drastic decrease in the number of people who are arrested for marijuana possession in the state um so clearly as paa's data shows uh the problem's not over and unfortunately the reality in Massachusetts what PA's map shows very clearly I think is that really only people of color are facing any kind of law enforcement interaction because of marijuana today in Massachusetts so DM had the effect of

49:39 making it possible for me to smoke weed and not have to deal with the police um but not for my black and brown neighbors so that's really the the issue that this map brings to the four and and so if you're uncomfortable with those racial disparities in marijuana you should vote Yes it's my view yes yep yes uh thanks for your work it's fantastic um I have two comments one black lives matter has looked at this issue but they're not narrowing it to Catabus use only number two uh the racialized and gendered aspects of this are really troubling uh people are not looking what's happening to black women okay they get criminalized at at higher rate even compared to black men uh so it would be helpful to take a look at that population segment of the population as well and finally has anybody looked at environmental justice there are environmental justice communities right here in Boston that no one has done anything about for decades thanks yeah thank you

50:43 cool so thank you for your talk um thanks my question is about um how do so you're doing work and I know you're doing good work and you're reputable and you're working with a very reputable organization so the graphs and charts I see that you're producing I can be personally trust as reliable um I'm wondering if there's any as we're getting into um an era where we're using a lot more sort of visualization data um if there's any kind of ways to evaluate whether something is um valid as opposed to propaganda say even if it looks valid and I mean we can just see like how people think people can spin things in all sorts of ways and um a lot of us might maybe can't necessarily look at the code and or anything so I'm wondering do you have any ideas or are there any organizations that will evaluate sort of the data that's being put out in visualization form to tell you if it's is trustworthy that's a

51:45 great question and I don't know how to answer and yeah we are um basically appealing to our reputations you know like so uh that's also why the code is there but I recognize that not all the people can read the code and um basically of course in the Bridgman litigation one of the arguments of the Das was that it was not PE reviewed like the my affidavit was not PE reviewed and yeah it is not right so um but that doesn't mean that the uh oh but to contact that I think what I'm doing is not an analysis like you know I I'm not trying to I'm not using uh AI or uh deep learning or anything like that I'm trying just to describe things and I try to keep it like like that it's like this table is brown and the computer is gray and if you put them together it's a you know

52:47 stuff like that it's just description MH that way I I I think I can convey the message very easily uh without trying to to uh to like involve more uh more thinking or more confusion so yeah I don't know it's it's so complex yeah but I think thanks for asking that yes oh sorry I'm not the same thank you so much this has been amazing umk I wonder it's clear that there's still a great deal of work valuable work to be done with this data set and forward to hearing more especially in light of next week's um Court proceedings but I'm wondering if you've identified like what's the next big data set you're going to go to work on that might produce marvelous results I'm sure there are many out there yeah one of them um like I I developed tools to explore data one of them is uh like a SBS like a p

53:50 pie chart uh What uh enable like in in how how do you say like an expanded uh pie chart you know like a Sunburst that's the name of it and what I use it for I can show you is um for instance with the payroll of Boston's um Boston city employees city employees yeah uh wait it's table yeah so we have I already imported the data sets and this is the payroll and these are the columns we have the name of the employee the department it's t uh their title and then we have all these uh income categories and the zip code so if we want to see divided by Department we can see that Boston Police Department

54:51 and Boston fire department have a big chunk chunk of the payroll like oh the employees these are only counting Counting employees but if we if you want to compare it to yeah in terms of money let's say oh total no wait I need to restart because okay um so we have this and then if we want to do total earnings and we say sum and it's money and we do this and now we have two this is the total earnings and this is the size of the department so you can see that uh like in proportion BPD is getting more money but if you want to divided by I don't know title it can get very complex well it's

55:53 a lot of data so you start to explore this type of things and now let's focus on I don't know uh can I one of the cool things that we saw with this tool is that for example Boston police officers who work in in West Roxberry get way more overtime than police in other parts of the city which is so weird it's like what's going on in West rockbury that all these cops have like huge hugely greater percentage of overtime than police elsewhere so that's the kind of you know unique Insight that these tools allow you to see that you wouldn't be able to see elsewhere no police off people who work there you get the data this is open data uh you have to go to each different municip yeah this is only for the seity of Boston open data yeah so now we have this which is uh Boston Police

56:54 Department only and then by title which is police officer and then by zp code so this enables you to see the the thing that K was saying which is that in this in these um zip codes like the police officers make significantly more money than in other C codes yeah and those are the neighborhoods where there's no marijuana yeah yes a about I really demonstated um what I mentioned before like even thinking about data is the the mission

57:56 like that I consider this being a mission oh sorry Mission driven data advocacy so that's how I see it it's like we have a mission which is prove that the uh inter the police interactions are racially biased and how do you which data do you need so you start to ask these questions through research and in many times you will find that there are government uh uh departments that have data a lot of data but but um and this is one problem I faced in my one one of my previous jobs which was uh in Mexico City's government that we had a lot of data but there was no impact of that data because there was no translation uh from that I I mean first they don't know um they don't know who is going to use it uh and second they don't actually uh do the Outreach for people to use it so it's it's like a

58:56 I don't know like a NE and chicken problem I think but if you put a mission on it and you can start to make all these diagnosis of oh I need this data set and and then I need to ask this person for what this colum means and then I need to do that but yeah it's it's very specific and very Mission specific things yeah it's there's no Explorer and there's no like uh it's complex thanks so much for the talk um I'm a journalist in South Africa and a lot of my work is bringing sort of information uh complicated information to sort of communities um I was wondering what was your interaction like with journalists did you bring them on early did you work with them in that way or was it a case of it's finished and then handing it over what was that sort of process like um we we haven't had that much interaction with journalists yet um because I mean we um so it's I love

59:59 to to have more more uh contact with them but it hasn't happened and I really don't know the ways to do it and uh also the the little contact we we've had is like they they won't publish the maps they are only going to write about it you know like especially here that it's my understanding that understanding that uh Boston's um media is very specific in this in its actions and how how they handle cases do they feel that they want to have had control over the whole process to publish it like it's almost like they don't want to collaborate is that feel I don't know uh I I I mean uh it's also through that I'm working on a an organization right so it's it it might be also that that they want to replicate my findings and they but they need the skills to do it so yeah it's it's like a catch 22 but I

01:01:01 love journalism and I think law and journalist are the the like the the key of our society but yeah that's great I think we'll probably leave it there join me and thanking PA for this terf talk thank you