I'm trying to plot a time series chart with Python, Matplotlib and Seaborn, but I can't have my xaxis with correct dates.
This is my df:
| Date | Hour | Rainfall |
|---|---|---|
| 2017-03-01 | 00:00:00 | 0.0 |
| 2017-03-01 | 00:15:00 | 0.0 |
| 2017-03-01 | 00:30:00 | 0.0 |
| 2017-03-01 | 00:45:00 | 0.0 |
| 2017-03-01 | 01:00:00 | 0.0 |
My data's date are measured every 15 minutes, and I used the code below to concatenate everything to a month series:
dfm = df.resample('M', on = 'Data').chuva.sum().reset_index()
dfm
And had this output:
| id | date | rainfall | mov_avg |
|---|---|---|---|
| 0 | 2010-04-30 | 99.0 | 99.000000 |
| 1 | 2010-05-31 | 111.0 | 105.000000 |
| 2 | 2010-06-30 | 111.8 | 107.266667 |
| 3 | 2010-07-31 | 347.8 | 167.400000 |
| 4 | 2010-08-31 | 66.8 | 147.280000 |
And just to check, all dates were converted with pd.to_datetime(). So, all of them are datetime64[ns] type.
I have already looked all these similar strackover flow sources to try to fix my problem, but I still couldn't find a solution:
(Editing the date formatting of x-axis tick labels in matplotlib)
(matplotlib - plot wrong datetime)
(Plotting dates on the x-axis with Python's matplotlib)
(Seaborn Plot - Wrong Dates on X Axis)
This is how I'm plotting my time series:
fix, ax = plt.subplots(figsize = (12,8))
sb.barplot(data = dfm, x = 'Data', y = 'chuva', ax = ax, color = 'blue')
ax.set(title = 'Time Series',
xlabel = 'Months',
ylabel = 'Rainfall (mm)')
#adding mov. avg.
dfm['mov_avg'].plot(kind = 'line', color = 'red', ax = ax)
# Define the date format
date_form = mdates.DateFormatter("%Y-%m-%d")
ax.xaxis.set_major_formatter(date_form)
ax.xaxis.set_major_locator(mdates.MonthLocator(interval=1))
ax.xaxis.set_minor_locator(mdates.MonthLocator(bymonthday=1))
#rotating labels
plt.xticks(rotation = 45)
;
Anybody has any idea of how to fix this?
