CorrelationBot is a simple way to find patterns in your data. It speaks HTTP, and has just one endpoint.
CorrelationBot speaks JSON over HTTP at http://correlationbot.com. Use GET for instructions, POST for results. Enjoy!
>>> import json
>>> import requests
>>> headers = {'Content-type': 'application/json', }
>>> resp = requests.post("http://correlationbot.com", headers=headers, data=json.dumps({
"data": [
[1, 2, 3],
[412, 5, 6],
[45, -125, 6.334],
# Or any number of equally-sized columns.
]
})
>>> print resp.json()
{
"correlations": [
{
"column_1": 1,
"column_2": 2,
"correlation": -0.86495821895559954,
"pearson": -0.86495821895559954,
# As you might guess, spearman, kendall, and others coming soon.
},
{
"column_1": 1,
"column_2": 3,
"correlation": -0.21695628247970969,
"pearson": -0.21695628247970969,
},
{
"column_1": 2,
"column_2": 3,
"correlation": 0.67754874098457696,
"pearson": 0.67754874098457696,
}
]
]
$ curl -X POST -H "Content-Type: application/json" -d '{"data": [[1,2,3,4,6,7,8,9],[2,4,6,8,10,12,13,15]]}' http://correlationbot.com
{"correlations": [{"pearson": 0.99535001355530017, "column_2": 2, "column_1": 1, "correlation": 0.99535001355530017}]}
CorrelationBot was made by Steven Skoczen in the GreenKahuna Skunkworks. Send us a note if you enjoy it! It source code lives on github, and is MIT licensed.