Emergent Algorithmic Intelligence Winter School 2023
The data can be downloaded here:
- train.pkl
- test.pkl
- test_full.pkl (Test with true output)
The pickle file data/prediction_example.pkl contains an example of a valid submission. Please provide your submission by mail in the same format. A single column with the same MultiIndex as the example file. The column name does not matter.
Variable | Description | Available Measurements |
---|---|---|
date |
Date of the measurement (hourly levels) | YYYY-MM-DD HH:MM:SS |
state |
The state where the measurement was taken | string |
city |
The city where the measurement was taken | string |
PM2.5_target |
Target variable (One hour ahead PM2.5) | ug/m3 |
PM2.5 |
Particulate matter 2.5 | ug/m3 |
CO |
Carbon monoxide | mg/m3 |
CO2 |
Carbon dioxide | mg/m3 |
NO |
Nitric oxide | ug/m3 |
NO2 |
Nitrogen dioxide | ug/m3 |
NOx |
Nitrogen oxides | ppb |
NH3 |
Ammonia | ug/m3 |
SO2 |
Sulfur dioxide | ug/m3 |
Temp |
Temperature | °C |
AT |
Air temperature | °C |
BP |
Barometric pressure | mmHg |
Benzene |
Concentration of Benzene in the air | ug/m3 |
CH4 |
Methane | ug/m3 |
Eth-Benzene |
Concentration of Ethylbenzene in the air | ug/m3 |
HCHO |
Formaldehyde | ug/m3 |
Hg |
Mercury | ug/m3 |
MH |
Mixing height | m |
MP-Xylene |
Concentration of Meta-Para-Xylene in the air | ug/m3 |
NMHC |
Non-methane hydrocarbons | ug/m3 |
O Xylene |
Concentration of Ortho-Xylene in the air | ug/m3 |
Ozone |
Ozone concentration | ug/m3 |
Power |
Power consumption | W |
RF |
Rainfall | mm |
RH |
Relative humidity | % |
SPM |
Suspended particulate matter | ug/m3 |
SR |
Solar radiation | W/m2 |
THC |
Total hydrocarbons | ug/m3 |
Toluene |
Concentration of Toluene in the air | ug/m3 |
VWS |
Vertical Wind speed | m/s |
WD |
Wind direction | degree |
WS |
Wind speed | m/s |
Xylene |
Concentration of Xylene in the air | ug/m3 |
The goal is to predict PM2.5_target
, the one-hour ahead PM2.5 concentration, on the test set.
This data is a modification of publicly available data. Please do not use the original data set for your solution.