Bangalore Traffic Open Data

[Mahesh Shantaram]

Hello all,

I wanted to share a small civic data project I've been quietly working on since last year...

https://github.com/thecont1/traffic-monitor-lizard

(the dataset)

This is a continuously-growing, open, versioned dataset of road traffic conditions across a set of pre-determined Bangalore routes, each roughly 10km long. The premise is simple – it asks Google Maps "how long do you think it takes to get from point A to point B?" The data has been accumulating twice an hour, every hour of every day for about 8 months now. The goal was to build a cheap-to-maintain, easy-to-manage, transparent, trustworthy, open, public traffic dataset for research and civic engagement. 

In the past month, I've also built an app on this dataset, which makes the data highly accessible and usable. https://co.thecontrarian.in/

trafficOracle lets you compare current travel speeds against historical patterns. You can even go back in time to see what the traffic was like on a given day at a certain time. It's a powerful dashboard with no backend, nothing to install, fully open source. I'd love for members here to poke around, raise issues, and more importantly, adapt it for other cities anywhere in the world.

Happy to answer any questions!

Regards,

Mahesh

__________

Mahesh Shantaram  |  www.thecontrarian.in  |  Bangalore  |  +91 99801 29770

[Mahesh Shantaram]

Hello all,

I wanted to share a small civic data project I've been quietly working on since last year. This is a continuously growing, open, versioned dataset of road traffic conditions across a set of pre-determined Bangalore routes, each roughly 10km long. The premise is simple – it asks Google Maps "how long do you think it takes to get from point A to point B?" 

The data has been accumulating twice an hour, every hour of every day, seven days a week for the past 8 months now. The goal was to build a cheap-to-maintain, easy-to-manage, transparent, trustworthy, open, public traffic dataset for research and civic engagement. So...

https://github.com/thecont1/traffic-monitor-lizard

In the past month, I've also built an app on this dataset, which makes the data highly accessible and usable. The app lets you compare current travel speeds against historical patterns. You can even go back in time to see what the traffic was like on a given day at a certain time. It's a powerful dashboard with no backend, nothing to install, fully open source:

trafficOracle
https://co.thecontrarian.in/

I'd love for datameetvadis to poke around, raise issues, and more importantly, adapt it for your city. Happy to answer any questions!

[Mahesh Shantaram]

Hi Sarath,

It costs ₹0 to keep traffiCOracle running because I scrape Google Maps directly via a headless browser. No API calls, zero backend. That's 13 Bangalore routes x 3 times an hour x 24 hours... I've not had any problems for the past 72,000 rows, but...

...to bring more cities into the fold, I ask that each city be adopted by a volunteer. It needs someone who knows a city well enough to curate a selection of routes, maintain the dataset, and most importantly, host the scraper on their own GitHub/Cloudflare/Vercel/Netlify/Gitlab, etc. etc. 

I've done quite a bit of testing up until now and can declare that for the most part it's a self-maintaining system. Just turn on the switch and forget about it! Again, total cost of operation = zero :)

Regards,
Mahesh

__________

Mahesh Shantaram | www.thecontrarian.in | Bangalore | +91 99801 29770

On Monday, 15 June 2026 at 18:23, Sarath Guttikunda <sgutt...@urbanemissions.info> wrote:

Good evening, this is very fascinating.

We used a similar method to extract speed information from the Google Distance API. When we conducted this study, Google provided free daily credits, which allowed us to operate within that margin. We successfully pulled speed information for a set number of routes across 20 Indian cities and a few other international cities.

https://urbanemissions.info/india-apna/

(top 20 cities)

This data reveals a great deal, highlighting traffic hotspots and the range of speeds that a city experiences. The data generally gives us a good understanding of the lay of the land. These data pulls also allowed us to capture the road lines. Consequently, we used that network as a proxy for the most heavily used roads -- which comes in handy for spatial allocation of estimated emissions in the airshed. This approach served as an excellent proxy to understand exactly where the majority of traffic moved, assuming drivers used Google Maps.

I am just wondering about your current cost element for this project. We eventually stopped running these queries because the API usage became extremely expensive.

(Gridded speeds were used in other emissions management exercises as well)

With best wishes,

Sarath

--
Dr. Sarath Guttikunda
https://www.urbanemissions.info

--
Datameet is a community of Data Science enthusiasts in India. Know more about us by visiting http://datameet.org
---
You received this message because you are subscribed to the Google Groups "datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datameet+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/datameet/CAJY8joTRrmkN5uTV%2BiSxAAbwHZbmk0BJ%3Dkrcc%2BCZa%2BjjCTtodQ%40mail.gmail.com.