realdata 0.1.1

Get real univariate time series data easily for testing

realdata

A mini-package for people who just want some real time-series data, any real data. Think of this as the alternative to

 import np
 x = np.cumsum(np.random.randn(1000))

Install

pip install realdata

Univariate live data (around 1000 points)

from realdata import get_live
values = get_live()

Multivariate historical data (around 30,000 points for 20 variables)

from realdata import get_historical
df = get_historical()

Alternatives

See other ways to grab the live data directly, or see this short tutorial on retrieving historical data using the microprediction package.

The historical data is from precisedata.

What's the data?

All sorts of things: traffic, electricity, emoji usage, hospital wait times, altitude of UFOs, security wait times at airports, changes in crypto-currencies, ... you get the idea.

Streams are live so the next time you get data, it will be different.

Can I add to the collection?

Yes see the tutorial on making a new stream of data.