Making data mean more through storytelling | Ben Wellington | TEDxBroadway

Translator: Timothy Duffy Reviewer: Denise RQ Hello everyone, as you say, I'm Ben Wellington and I'm a

data

storyteller. A

data

storyteller. If you had asked me a year ago what a data storyteller was, I would probably say I have no idea. So today I'm going to tell you about my journey over the past year, where I accidentally became a data storyteller. I'll tell you what I've learned along the way and maybe convince some of you that you too can be data storytellers if you're curious and willing. A little about my background. First, I work at an investment and technology company called Two Sigma, where I do data science.

So that's kind of part of my world. But I also married an urban planner, so I have my IT world and my urban planning world, and for most of the time, those two sections of my life have been pretty separate, and that's how it was. Until something interesting happened here in New York, in 2011, then-Mayor Bloomberg signed this legislation called Open Data Laws in New York. Open Data Laws are really interesting to people like me because they take data that is within the city government and suddenly allow anyone to see it. Whereas before the government would look at something and tell us, "Hey, this neighborhood has so many accidents," now we can see point-by-point data of what's happening at a very local level.

When these two things came together, they ended up... by the way, there was an open data portal that I should point out, that anyone can go to, it's the New York open data portal, and there are data sets on all kinds of things . In fact, there's one about the size of the TVs in Times Square and their location. I don't know what to make of it, but it's really cool. There are data sets of all different types; in fact, there are

more

than 1,200 data sets so far. And it's growing all the time. I kind of took this data science work and my interest in urban planning from conversations with my wife and put it together into this blog, called I QUANT NY.

Awesome thank you. One of the first things I made was this map. And this is a map of cycling injuries in New York City. (Laughs) Hilarious injuries. Red areas are areas where people were experiencing

more

cyclist accidents. Then I found this through some public data and mapped it all out. I notice a few things: one, that on the East Side of Manhattan there were more bicycle accidents, more injuries, because that's where there are more cyclists getting off the bridges. But there were also other hot spots, like Williamsburg in Brooklyn or Roosevelt Avenue in Queens. I wrote about it and blogged about it, and to me it was fairer to learn how to make maps;

It was this open source software called QGIS, I wanted to learn it and when I did, something interesting happened, people started writing about it. Gothamist covered it and Brokelyn claimed it was a "death trap," which isn't exactly what I said. (Laughs) Streets Blog, and then even in The Atlantic. This is kind of from a blog I put up on Tumblr and I had no followers, and that was really interesting. Eventually I started saying, "Why are people writing about that? I'm not the first to analyze cyclist accidents or make maps like this." What I did wasn't that complicated, what made it spread?

And I thought about it and worked on the next few posts, and I could see what was moving and I realized that there was a third part of this that was really, really important. I

mean

, you probably won't see it coming: improvisational comedy. Yes, improvisational comedy. I've been improvising since I was at a summer camp called French Woods. Yeah! French Woods Students! - in upstate New York. I've been improvising since I was 13 and I've been doing it ever since. I learned a lot of things in improvisation and I realized that I was incorporating them into my writing, into this data science so that people would be more interested in it.

I believe that to spread science you need to be able to tell stories, so I'm going to tell you why improv relates to data science and how you can come together to tell better stories. That's why I call it

storytelling

with data. First, in stories you want to connect with people's experiences, right? If you're doing an improv scene, you learn that if you brush your teeth next to your wife, something that people can relate to, what they look like when they brush their teeth, you can make a scene through that. People identify with it because they have experiences.

I tried to write about things that New Yorkers experience and I thought, "What do New Yorkers experience more than Duane Reade?" (Laughs) Right? I thought this might be interesting and it turns out it was. I mapped every building in New York to the nearest pharmacy and colored it next to the pharmacy. Orange is Duane Reade, red is CVS, blue is Walgreens, and yellow is Rite Aid. First of all, this neighborhood is Duane Reade country. No doubt. I also learned that CVS and Rite Aid are attacking from the water. (Laughs) Good strategy CVS, coming from the Hudson.

Duane Reade won't see you coming. (laughs) I thought, "Is Duane Reade really a New York thing?" Turns out no, it's a Manhattan thing. Zoom out even further, and the Bronx is Rite Aid country. Brooklyn and Queens are a mosaic. If you work for one of those companies it's probably very interesting. For me, it's interesting to just ask ourselves what our experiences are and how we can quantify them and tell stories from our lives. That's something we can all relate to and is part of

storytelling

, even in data analysis. I noticed that you want to focus on a single idea.

In an improv scene, you can try to have seven ideas going, but things can get lost pretty quickly. In my job I try to focus on one idea and I went and looked at the Citi Bike data. It's interesting, there are people who leave and come to the stations, there is a lot of data. What happens if we take an idea and that idea is gender? Here, I mapped the percentage of male and female Citi Bike riders in New York. What we can see is that in this neighborhood more than 80% of the cyclists are men. This is a very predominantly male Citi Bike neighborhood.

What does that tell us? It could be our transportation infrastructure. It could be a gender study here in the city. Also, if you're looking to meet a girl on a Citi Bike, go to Brooklyn, that's important. (Laughs) The important thing here is that it's an idea, it's just gender. There are many columns, it is a large data set, but let's study just one. The other thing is to keep it simple. Not just an idea, but a simple idea. The ideas can be very complex and you find yourself improvising thinking, "This guy goes here, and this, and this," and you'll lose everyone very quickly.

That's why I also try to keep it simple. So when people hear that I do math, they often think I do this. But it's more like that. (laughs) I

mean

, I just tell things. Sumo, maybe I make a percentage. This is all just high school math, not crazy math. People can really do this if they stop and start asking questions. An example: I looked at the percentage of out-of-state license plate parking tickets in each borough of New York. We see that in this neighborhood there is a higher percentage of people who come from out of state and receive tickets.

Which is telling; People are more likely to drive into Midtown, and as we get farther out in our districts, there are fewer commuters, which also makes sense. I also wanted to do this by state. First, I did New Jersey. Midtown, yes, the data shows that people from New Jersey drive to Midtown. Connecticut, a completely different panorama. Coming from the north. You can actually see it. They go to the Bronx Botanical Garden. And my absolute favorites, Californians, where do you think they hang out? The trendiest areas of Brooklyn: Williamsburg, Bushwick, Green Point, that's where Californians get parking tickets.

It can tell us a lot about our city by looking at our data. Also, explore the things you know best. You all come from different fields. You know the things you know. You know very well the area you study, the area in which you work. I'm learning New York, I've lived here for over a decade, so I'm focusing on New York. In an improv, if you are a lawyer and you go into a scene as a lawyer, that scene will be good, because you know all the vocabulary. You can play it to the top of your intelligence and just hit it.

I tried other cities, but it's difficult because I don't have any context. For today I did an analysis of Times Square and thought, "What can we all relate to?" Maybe take a taxi. I was curious to know where people hail taxis around Times Square. That's 8th Avenue around here, you can see that people generally take taxis on 8th Avenue when they leave this district. Not so far south on 7th Ave. You can see the big yellow blurb on the bottom left is the Port Authority, so it makes sense. But people are really heading towards the avenues.

The interesting thing here is that this is where people take taxis. Where do you think people get out of taxis? They come out more on cross streets. It's much more of a grid. If you're taking a taxi somewhere, especially if you're a tourist, you give the person the address and they take you there. They won't leave you on the corner and say, "Good luck, buddy." That doesn't happen. What I really liked about this was that depending on the direction of the street, people got out of taxis differently. There is 7th Avenue: if you enter on 46th, you go from west to east, it seems that you get off on the west side of the avenue.

Because? If a taxi ever gets stuck in traffic, just get out. In fact, you can see it in the data; If you enter from the other side, you can see people coming down that side of the avenue. Depending on the direction of the street, people get out of taxis because they are probably waiting at traffic lights. And this is interesting, right? If you advertise in the district, you may want to know where to welcome them and where they hang out most often. We can start studying this with our public data. With this, you want to try to make an impact.

I tried to make it in city government by doing some of this work. Each of you has your own ways of

making

an impact. In particular, I made a proof, a mathematical proof, that no matter how many times you ride the subway and refill using its buttons, you cannot get a balance of $0.00. (Laughs) Literally, it's not just you, or you, or you. You literally cannot get a $0.00 balance if you use their buttons. There is a trick: you can write $19.05 and get a balance. When I wrote about this, the MTA responded. I said, okay, maybe this is a shock.

And they said: "These machines do not contain an infinite amount of change and the denominations are suggested to ensure that there is enough change to accommodate cash paying customers. That said, we will certainly consider this as part of the process involved in implementing the next scheduled rate increase." (Laughter) So, the rate increase is coming. Imagine in March, when you say, “I want $20 on a MetroCard,” and you get one. "How much would that be?" And they tell you: "$18.43" and you pay. Unlike now, "I'm going to pay $20." "We will give you a random amount above that amount." Imagine if we changed that, we could run our city better.

We'll see if the MTA delivers, I'd love to have an impact there. I also found something strange: in half of the taxis in the city the tip is based solely on the price and the surcharge. So if you get into a Verifone taxi and press the 20% button, you are actually paying 20% ​​on top of the taxi fare and a small surcharge. If it is managed by Creative Mobile Technology, the other half of the taxis, and you press the 20% button, you are also paying taxes and tolls. So for two different computers, you are paying more tips on one of the computer settings than the other, because it is calculating the tip in addition to the tolls.

Is this all a big deal? Well, those drivers are

making

$250 more a year in tips, with this little rounding. We have half of our taxis where we all pay a little more, and the drivers too, which is not bad but it is somewhat inequitable. When I pointed this out to TLC, they said, "We appreciate the work put into this analysis and are reading it carefully." (Laughs) Impact. (Laughs) I'm working on it. And my favorite was this: I mapped fire hydrants in New York City. These are not simply fire hydrants. These are the fire hydrants assigned by the amount of parking ticket revenue they are generating.

These are the top 250 culprits in New York. First of all, be careful on the Upper East Side, District 19 will fine you no matter where you park to put up a hydrant. More interesting were these two fire hydrants that were installed on the Lower East Side and generated $55,000 a year in fines. Two hydrants, $55,000 for about 5 or 6 years. Finally the data is public, I took a look and when I went to find out what was going on, it turns out there's basically a fire hydrant, then a bike lane, and then a parking spot. So you're thinkingthat you are not in front of the sidewalk or the hydrant, that there is a bike lane between us.

It turns out that while the DOT painted over a parking spot, the NYPD disagreed. So they would fine the place for years and years. This is actually a shot of the Google StreetView car that drove by and captured the ticket, which I really appreciated. I wrote about it and heard again from the city: "While the DOT has not received any complaints about this location, we will review the road markings and make appropriate modifications." That is a statement of action, we are improving. I thought, "Well, Government, I'll keep trying to make an impact." And suddenly they repainted the place.

Yeah! Impact! (Applause) Someone's listening, this is great. You can look at your data and you can have an impact. Sometimes your message doesn't get through, but where I know a statement was made, I know that at least I've changed some thinking in these agencies, that someone thinks about those things. I think those are also having an impact. To do that, again, I think you really need to think about storytelling, like connecting with people, trying to get an idea across, keeping it simple, and exploring the things you know best. In case you think this whole data thing is not for me, but for IT people, the open data portal is easy to use.

I teach a statistics class at Pratt, for urban planners, and on our second day of class one of my students turned this, which is a list of accidents in the neighborhood around Pratt, into this analysis of injuries by vehicle type. And this is just one or two classes that use Excel, this is not crazy programming. If that scares you, here's a dataset of graffiti in 311 complaints. And a student in a program called City Term, which takes students to learn about the city, it's a high school semester program, a student named Abby created this map. So if you are afraid of computers, that's okay too.

Understanding where graffiti is in New York. It's not about being a computer scientist, anyone can collect data, you just have to know what questions to ask and try to tell your own story. I just hope everyone realizes they can tell stories with data. Thank you. (Applause)