Ambient particulate air pollution and daily stock market returns and volatility in 47 cities worldwide

Major stock index data collected from Investing.com. After identifying the coordinate points for all the corresponding stock exchanges, air pollution and meteorological data was collected from as near as possible. All meteorological data was downloaded from the NOAA GSOD database, and air pollution data was collected from various sources that are described more accurately in the metadata file. Sufficient daily data for fine particulate matter concentrations were found for 47 of the locations.

The meteorological variables contained a rather large amount of randomly missing observations which were imputed using CART (Classification and Regression Training) method from MICE package for R.

Variables in masterdata:
log_ret - logarithmic daily returns
pm25 - daily PM2.5 concentration in ug/m3
lag1_ret - 1 day lagged returns
lag2_ret - 2 day lagged returns
temp_bins - temperature as 2.5C wide dummy bins
dewp_bins - dew point as 2.5C wide dummy bins
RH - relative humidity
DEWP - dew point
TEMP - temperature (C)
SLP - sea level pressure
VISIB - visibility
WDSP - wind speed
PRCP - precipitation
woy - week of year
dow - day of week
vixopen - daily VIX index value
date - date
city - name of the city