(df.index)): TypeError: '>' not supported between instances of 'str' and 'float'

Hi Jonas, Thanks for this wonderful gem of a library 😃 HERE is a comment from you a while back. Now I’m trying to plot data from a json object.

THIS Is the API endpoint.

When I try below code based on your advice (I’m nearly certain I’m doing it wrong),

import requests as rq
import pandas as pd
import finplot as fplt


url = 'https://www.alphavantage.co/query?function=TIME_SERIES_INTRADAY&symbol=IBM&interval=5min&apikey=demo'

r = rq.get(url).json()

ax = fplt.create_plot('test', rows=1)

# plot candle sticks
fplt.candlestick_ochl(r['Time Series (5min)'], ax=ax)

I’m met with this traceback,

Traceback (most recent call last):
  File "C:\BeingProfessional\projects\Py\SoFlow.py", line 22, in <module>
    fplt.candlestick_ochl(r['Time Series (5min)'], ax=ax)
  File "C:\Program Files\Python37-32\lib\site-packages\finplot\__init__.py", line 1254, in candlestick_ochl
    datasrc = _create_datasrc(ax, datasrc)
  File "C:\Program Files\Python37-32\lib\site-packages\finplot\__init__.py", line 1819, in _create_datasrc
    datasrc = do_create(*args)
  File "C:\Program Files\Python37-32\lib\site-packages\finplot\__init__.py", line 1818, in do_create
    return PandasDataSource(pd.concat(args, axis=1))
  File "C:\Program Files\Python37-32\lib\site-packages\finplot\__init__.py", line 160, in __init__
    if type(df.index) == pd.DatetimeIndex or df.index[-1]>1e7 or '.RangeIndex' not in str(type(df.index)):
TypeError: '>' not supported between instances of 'str' and 'float'

My queries are,

1. How do I plot this json response data in finplot? If it makes it easier, I am okay with a solution with pandas.
2. Why does finplot throw this error?
3. If there is something wrong with this json formatting, how about the TD Ameritrade API’s price history json? I know how to use TDA API, but how do I plot that data in finplot? The json schema looks like THIS

I’ll appreciate any help or suggestion…

[highfestiva]

Btw. this is what you need to do to get your first example plotting:

url = 'https://www.alphavantage.co/query?function=TIME_SERIES_INTRADAY&symbol=IBM&interval=5min&apikey=demo'
r = rq.get(url).json()
ts = r['Time Series (5min)']
rows = [[t]+list(kv.values()) for t,kv in ts.items()]
df = pd.DataFrame(data=rows, columns=['timestamp']+list(next(iter(ts.values()))))
df = df.rename(columns={'1. open':'open', '2. high':'high', '3. low':'low', '4. close':'close', '5. volume':'volume'})
df['timestamp'] = pd.to_datetime(df['timestamp'])
df = df.set_index('timestamp')
df = df.astype(float)

Note though, that the open and close columns are actually inversed in the downloaded data! We fix by simply flipping open and close when plotting like so:

fplt.candlestick_ochl(df[['close','open','high','low']]) # close before open
fplt.show()

[highfestiva]

… regarding your first example/bug you need some more data processing, but out of the scope for a plotting library:

1. Put each timestamp per row.
2. Make it into a matrix.
3. Convert all values from strings to their respective types (datetime and floats).

Then you can plot.