Y Axis Values Out of Order on Plotly Graph Python

📊 Plotly Python
1

I am using AlphaVantage API to download data points, which I then convert to a pandas DataFrame.

I want to plot the new DataFrame with a Scatter/Line graph using Plotly. The graphs seems to come out perfect when plotting them in Google Colab, however, I seem to be unable to replicate my results in PyCharm and Jupiter Notebook.

When plotting in either PyCharm and JN, the Y-Axis values are plotted out of order, like the graph is trying to create as straight of a line as possible (see 2nd image and look closely at y-axis).

Here is a simplified example of the code and graphs (desired outcome vs problem graph):

Exact Same code used in both instances

Screen Shot 2021-06-11 at 10.10.15 PM
Screen Shot 2021-06-11 at 10.10.15 PM1717×445 51.2 KB

Screen Shot 2021-06-11 at 10.10.02 PM
Screen Shot 2021-06-11 at 10.10.02 PM1876×900 40.9 KB 

import requests
import pandas as pd
import plotly.graph_objects as go


# DATA FROM API
response = requests.get(url='https://www.alphavantage.co/query?function=TIME_SERIES_WEEKLY&symbol=IBM&apikey=demo')
response.raise_for_status()
stock_weekly = response.json()['Weekly Time Series']


# CHANGE DATA FORMAT, RENAME COLUMNS AND CONVERT TO DATETIME, FINALLY FLIP TO HAVE DATE IN ASCENDING ORDER
raw_weekly_data = pd.DataFrame(stock_weekly)
weekly_data = raw_weekly_data.T
weekly_data.reset_index(level=0, inplace=True)
weekly_data.rename(columns={
    'index': 'DATE', 
    '1. open': 'OPEN', 
    '2. high': 'HIGH',
    '3. low': 'LOW',
    '4. close': 'CLOSE',
    '5. volume': 'VOLUME'
    }, 
    inplace=True)
weekly_data['DATE'] = pd.to_datetime(weekly_data['DATE'])
weekly_data = weekly_data[::-1]
weekly_data = weekly_data.reset_index(drop=True)


# PLOT
fig = go.Figure()

fig.add_trace(go.Scatter(
    x=weekly_data['DATE'], 
    y=weekly_data['CLOSE']))

fig.show()

Thank you in advance!!!

2

Welcome to the Dash Community :slightly_smiling_face:

The error is due to the data types of your columns in your dataframe.
dtypes
The values are of type object.

However this was not a problem is previous versions of plotly (which must be installed on your Google Collab).
The newer releases requires the values to be numeric.

You can either convert the columns to numeric like this:

#converting to type numeric
cols = weekly_data.columns.drop('DATE')
weekly_data[cols] = weekly_data[cols].apply(pd.to_numeric)

or just add autotypenumbers=‘convert types’ to update your figure:

fig = go.Figure()

fig.add_trace(go.Scatter(
    x=weekly_data['DATE'], 
    y=weekly_data['CLOSE']))

fig.update_layout(autotypenumbers='convert types')

fig.show()
3

You’re a life saver!

I used a combination of both and it gave me exactly what I was looking for. Thank you so much

1 Like