search

OliverSherouse / wbdata

A python library for accessing world bank data
GNU General Public License v2.0
182 stars 55 forks source link

Submit Pandas pull request? #1

Closed davidrpugh closed 10 years ago

davidrpugh commented 11 years ago

Have you thought about incorporating this into the Pandas project?

OliverSherouse commented 11 years ago

I honestly hadn't considered it. As I understand, the point of Pandas is data analysis, not retrieval, but I guess there's no reason not to ask if they're interested.

davidrpugh commented 11 years ago

Pandas has built-in functions get_data_fred() and get_data_yahoo() that will grab data from those sources and return a Pandas DataFrame. Extending the functionality of Pandas to include World Bank data makes sense.

I would also suggest including the option to return a Pandas Panel object instead of a DataFrame object. I suppose a country-year panel would be best with major axis being country and minor axis being time? I am a Pandas newbie, but if I figure out how to make this work within wbdata I will submit a pull request.

OliverSherouse commented 11 years ago

I'll take a look at that with Pandas. For the panel object, it should be as simple as wrapping get_dataframe, doing a multi-index, and creating a panel from that. I'll take a look at that as well.

OliverSherouse commented 11 years ago

To give an update on this, I've opened a new issue for the Panel request, and the code for that is basically written. I'll put further updates on that there.

As to pandas integration, it seems to me that the best thing will be a simplified function that replicates some of the functionality here, but not all of it. That matches their behavior with fred, for example. I will probably do that once I fix my code for Fetcher here, which ain't particularly pretty.

OliverSherouse commented 10 years ago

Going to close this bug as pandas seems to have taken care of it themselves (see http://pandas.pydata.org/pandas-docs/stable/remote_data.html).