Visualizing large scale Uber Movement Data

By Visit profile

New York's cab data visualization from Uber's Engineering blog
Last month one of my acquaintances in LinkedIn pointed me to a very interesting dataset. Uber's Movement Dataset. It was fascinating to explore their awesome GUI and to play with the data. However, their UI for exploring the dataset leaves much more to be desired, especially the fact that we always have to specify source and destination to get relevant data and can't play with the whole dataset. Another limitation also was, the dataset doesn't include any time component. Which immediately threw out a lot of things I wanted to explore.
When I started looking out if there is another publicly available dataset, I found one at Kaggle. And then quite a few more at Kaggle. But none of them seemed official, and then I found one released by NYC - TLC which looked pretty official and I was hooked.
To explore the data I wanted to try out OmniSci. I recently saw a video of a talk at jupytercon by Randy Zwitch where he goes through a demo of exploring an NYC Cab dataset using OmniSci. And since my dataset was very similar to that, I thought of giving it a try.
You can find the Jupyter Notebook here:
Just as a toy experiment I tried to answer and visualize the following.

And also a bunch of interesting facts we can glean from this dataset.
However, while trying to do this, I realized it's pretty hard to work on a huge dataset in jupyter directly if you load the whole dataset into a dataframe anyway.
I used OmniSci's Cloud interface to load up my data and then connect to that dataset using pymapd to read the sql data.
What I did not do was to be smart and utilize OmniSci's super powerful mapd core and slice and dice the dataset in the cloud itself. Which cloud have saved me a lot of time. For example, the query I was running on one-sixth of the whole dataset was taking 25 minutes.
You can take a look at some of my rough ideas, tries and more graphs in this Jupyter Notebook.
However, it seems OmniSci also has a super helpful visualization web interface as part of OmniSci Cloud called Immerse. And I was able to cook up these dashboards in less than 5 minutes.
And Immerse was able to crunch through the whole dataset (not one-sixth) almost instantly and produce these charts for me. I am pretty impressed with it so far. And it seems with help of pymapd and crafting some sql queries, I should be able to harness this speed as well. 
That would be my next try probably.

What's Next:

Since I realized how powerful OmniSci Immerse can be and starting to play with pymapd. My next pet project is merging Uber Movement's yearly data with ward based time series data. So that we can recreate the whole dataset and analyze some of the interesting aspects of it as we did above. I am mostly interested to see (preferably in Bengaluru data)
  • Uber's growth through time (and specific activity growth in different wards)
  • Figuring from historical time series which wards and routes have most traffic in which hour (this also should let us predict which areas may face surge pricing)
  • See if the growth has saturated in any specific place (should give us upper threshold for that area)
  • If an increase in Uber Demand directly co-relate to travel time (maybe the increased demand is causing traffic?)
  • Can we load it up in (more specifically using this demo as a template) and have a nice timeseries visualization?
I have a thing for Hackathon. I am a procrastinator. A lazy and procrastinator graduate student, not a nice combination to have. But still when I see hundreds of sharp minds in a room scrabbling over idea, hungry to build and prototype their idea. Bring it to life, it finally pushes me to activity, makes me productive.  That is why I love Hackathon, that is why I love HackRice, our resident Hackathon of Rice University.

TL;DR: if you just want to try the extension, chrome version is here and Firefox version is here.

I have been participating at HackRice since 2014, when I think for the first time it was open for non-rice students, and have been participating ever since. What a roller coaster ride it has been, but that is a story for another day. HackRice 7.5 being the last one I will be able to attend at Rice, it was somewhat special and emotional for me.

HackRice 7.5 was a tad different form the other iterations. For starters it was the first time it was being held in Spring semester…

Returning to my cubical holding a hot cup of coffee and with a head loaded with frustration and panic over a system codebase that I managed to break with no sufficient time to fix it before the next morning. 

This was at IBM, New York where I was interning and working on the TJ Watson project. I returned back to my desk, turned on my dual monitors, started reading some blogs and engaging on Mozilla IRC (a new found and pretty short lived hobby). Just a few days before that, FirefoxOS was launched in India in the form of an Intex phone with a $35 price tag. It was making waves all around, because of its hefty price and poor performance . The OS struggle was showing up in the super low cost hardware. I was personally furious about some of the shortcomings, primarily the keyboard which at that time didn’t support prediction in any language other than English and also did not learn new words. Coincidentally, I came upon Dietrich Ayala in the FirefoxOS IRC channel, who at that time was a P…

Reading Time: 7 MIn Some of you know I have been recently experimenting a bit more with WebXR than a WebVR and when we talk about mobile Mixed Reality, ARkit and ARCore is something which plays a pivotal role to map and understand the environment inside our applications. I am planning to write a series of blog posts on how you can start developing WebXR applications now and play with them starting with the basics and then going on to using different features of it. But before that, I planned to pen down this series of how actually the "world mapping" works in arcore and arkit. So that we have a better understanding of the Mixed Reality capabilities of the devices we will be working with.

Mapping: feature detection and anchors Creating apps that work seamlessly with arcore/kit requires a little bit of knowledge about the algorithms that work in the back and that involves knowing about Anchors. What are anchors: Anchors are your virtual markers in the real world. As a develope…