Thinknum: a new web platform for Financial analysis

[Justin Zhen]

Hi guys, I'm a co-founder at Thinknum. Thinknum is a tool for investors to collaborate on financial analysis. We are indexing all of the world's financial data and making it available on an open platform. With Thinknum, you can review companies and create your own valuations using our Cashflow Model. You can also research and manipulate financial data with our Time-series Analysis tool.

A few experts in our space have written about Thinknum:

• Francis Smart discussed Thinknum's integration with R on R-Bloggers.
• Dr. Gray wrote about our cashflow models at TurnkeyAnalyst.

Check us out at thinknum.com and please contact me at jzhen@thinknum with any feedback that you have. Thanks.

 

[Ian Kaplan]

The good part: the financials tab gives good information from the Corporate 10Q quarterly filing.
The no so good part: looking at the API it's not clear how to download the financials in, say, .csv form. An interactive interface is pretty much useless for anyone doing quantitative modeling where many stocks will be looked at.

 

[Gregory Ugwi]

Hi Ian,
Thanks for reviewing Thinknum. I work with Justin at Thinknum.

Downloading the data as csv is straightforward: for example is http://www.thinknum.com/api/v1/?expression=goog@Revenues.q&type=csv

will get you the revenue of Google.

Users typically request the data as json, however one can set the url type parameter to csv. In short all 7 million data-series thinknum indexes is available through our api.

Users such as Fidelity's WealthLab, a backtesting tool, and Spearian, currently access Thinknum's data programmatically.

As a former quant, I share your sentiments about interactive data being useless if unavailable though a straightforward API. This is exactly Thinknum's raison d'etre. The web provides a wealth of interactive data. we are here to make it useable for analysis.

we will work on improving our API documentation. Thanks again for the feedback.
Greg

 

[Ian Kaplan]

Thanks for the post, Greg.

Is it possible to download the complete set of quarterly values as a .csv table (or JSON object)? I would like to download all of the quarterly data and then store it locally to build factor models for portfolio construction.

I use R for modeling, so I tend to use .csv data (I think there's a way to get JSON data into R too, but I have not used it). If I were using Java JSON might be a better choice.

Does the quarterly corporate data that Thinknum is publishing come from the SEC XRBL filings (10Q reports in XML format)?

Best,

Ian

 

[Gregory Ugwi]

Hi Ian,
We have an R API on CRAN, Francis Smart had a nice description of this API on R-Bloggers.com

- The API currently does not support downloading a dataset in single call. Implementing this feature is high on our priority list.

- Yes, corporate data comes from SEC XBRL Filings. you can click on the labels of the filings on our site to pull up the time-series and you can click on a bar (in the bar chart visualization) to pull up the exact SEC filing.

Regards,
Greg

 

[Lucas Hawk]

I am trying to use ThinkNum for fairly simple data (equity prices, volume, etc). I am mainly hoping to use it as a way to sort of aggregate data sources in R.
However, the documentation seems to be completely non-existent. The CRAN documentation is almost empty, and the ThinkNum website appears to be still under construction. (http://www.thinknum.com)
I am guessing this is all still a work in progress, but do you have any documentation I can use to see the syntax, and get data?

 

[MRoss]

I could see MBA programs eating this up (if it ever took off in the financial industry). Well done guys.