I am getting wait times across a variety of geographic locations. The times are stored in a mongodb collection as a UTC timestamp, and when it comes time to present them, I convert them to local time. The dataframes come out ok, as in the data makes sense. However, the graphs seem offset by 2 hours.
Example of the graph I am illustrating
Example of the dataframe that feeds it:
Specific example of the offset error:
My code:
hist = pd.DataFrame(mongo_queries.get_hist_data(f)) # get the data
hist['local_time'] = hist.apply(get_local,axis=1) # convert UTC to local time
alt_color = {'Maximum':'max(wait):Q', 'Average':'average(wait):Q','Median':'median(wait):Q'}
hist
st.write('### Schedule View')
scatter = alt.Chart(hist).mark_rect().encode(
alt.Y('hours(local_time):O', title='hour of day',),
alt.X('day(local_time):O', title='Weekday'),
alt.Tooltip(['hours(local_time):O','day(local_time):O',alt_color[metric]]),
color=alt_color[metric],
).interactive()
st.altair_chart(scatter, use_container_width=True)
As you can see in the data, you would expect that the latest times (within the 7th hour AM) would bin together and average to 17.75. The average is correct, but it is being illustrated on the 9AM block…
After some googling it seemed that the date format was causing this issue, so I converted the local_time column to an ISOFORMAT which was apparently the fix, but it is not fixing it.
Upon further investigation, it seems that altair is converting it to MY local time, regardless. I put the UTC time as the input to the graph(alt.Y and alt.X above) and its still binning them to the 9th hour, which is what I am in currently…
I guess the question is, how do I stop Altair(I think that's the culprit) from converting times and letting them be on the specified local times. I am expecting each location to follow the same generic trends throughout the day, respective of their local times. With locations across the continent, if they are all mashed together in my timezone, it’s not going to work…
Another weird thing is, my UTC time seems to be the right time, but the wrong timezone offset. The local time is correct. I am confused…
