Data Provider Recommendation

Corey · Post by **Corey** » Wed Sep 12, 2012 4:48 pm

So I will preemptively apologize for this question. I tried to do a search and glean as much as I could, but I am a bit overwhelmed.

We are looking for a data-provider solution (or a set of solutions) that could achieve the following:

- Historical fundamental corporate information

- Historical economic information

- Historical derivatives information (multi-asset [equities, commodities, rates, fx] / consensus "index" data is fine for things like CDSs instead of individual trades). Bonus for pre-computed relevant greeks

- Historical corporate fixed income data

- Historical index constituent information

Mostly, we just need EOD data. By historical, obviously more is better -- but I know that gets very expensive for fundamental (and other obscure) data.

It would be a huge bonus if we can pull international data (e.g. foreign equity information) and point-in-time data (un-restated). Consensus information is great too (though realized is an absolute must). Again, these are bonuses, and we could likely live without them if they are unavailable / extremely expensive.

Ideally, we'd like to access the data via a server API call (e.g. "request to pull" instead of being "pushed" the data every night and having to maintain our own database, which is a total pain in the ass and, frankly, we don't have the tech staff for).

We do most of our development in Python on Linux -- so huge bonus points if an existing python client exists. If not, the more "REST"ful the API, the better. We'd survive if we had to interface through an Excel spreadsheet ... but we'd prefer not to.

So far, we're looking into services from Bloomberg, FactSet, CapitalIQ (CompuStat), Thomson Reuters, MSCI, but it is hard to determine if we'd be better off trying to just get one service or a mish-mosh of multiple services.

Any thoughts?

Thanks!

jungle · Post by **jungle** » Thu Sep 13, 2012 3:35 am

Datastream ticks several of your boxes (economic information, corporate bond data, index constituents, FX and rate vol, foreign equity data, point-in-time database, API). It's good for economic data, but I can't comment on equities/credit. Their main competitor for economic data is Haver Analytics, which might also be worth a look.

EDIT: Point-in-time databases for economic data are expensive and I know Bloomberg doesn't have this (at least as of May).

Trev · Post by **Trev** » Fri Sep 14, 2012 4:30 am

In regards to what Jungle was referring to, the St. Louis Fed has a economic database called FRED with a point-in-time sub-database called ALFRED. More than likely it doesn't have the type of API-based interface Corey is looking for but I thought I'd mention it.

Cheers.

briant57 · Post by **briant57** » Fri Sep 14, 2012 1:42 pm

Corey,

For Historical Economic data, do check out FRED as Trev mentioned.

There is a great api available, and if you are using python with the PANDAS module check out the pandas.io.data DataReader. It is v simple.

Corey · Post by **Corey** » Fri Sep 14, 2012 6:25 pm

Wow, don't know how I didn't see FRED before! Looks like it will cover the majority of my cases.

Since I already have a historic equity price provider, I'm really just looking for derivatives data (single instrument and index data for both exchange traded and OTC) and index constituent info. Any thoughts on the best provider for this single case?

Thanks for the pointer towards datastream, jungle, I'll check it out.

I use pandas, so thanks for the tip briant57. Must've missed that DataReader some how...

briant57 · Post by **briant57** » Fri Sep 14, 2012 10:34 pm

to get you started:

from pandas.io.data import DataReader

DF=DataReader('SP500', 'fred', '1/1/2008')