How can I generate a chart getting data from a JSON file-CodePudding

I have a json file in which I have different data, I need to generate a graph whose legend is "Dates" and whose data is all the viewers of each user. I've been trying to do it all morning, but I have no idea, can someone help me?

{
    "Avg": 49,
    "Max": 70,
    "Dates": ["01/01/2019 01:05", "01/01/2019 01:10", "01/01/2019 01:15", "01/01/2019 01:20", "01/01/2019 01:25", "01/01/2019 01:30", "01/01/2019 01:35"],
    "Viewers": [
        {
            "Name": "User1",
            "Viewers": [1,2,3,4,5,6,7]
        },
        {
            "Name": "User2",
            "Viewers": [2,3,4,5,6,7,8]
        },
        {
            "Name": "User3",
            "Viewers": [3,4,5,6,7,8,9]
        },
        {
            "Name": "User4",
            "Viewers": [4,5,6,7,8,9,10]
        },
        {
            "Name": "user5",
            "Viewers": [5,6,7,8,9,10,11]
        },
        {
            "Name": "User6",
            "Viewers": [6,7,8,9,10,11,12]
        },
        {
            "Name": "User7",
            "Viewers": [7,8,9,10,11,12,13]
        },
        {
            "Name": "Total",
            "Viewers": [28, 35, 42, 49, 56, 63, 70]
        }
    ]
}

CodePudding user response：

Assume that you have put the json data in a faile named input.json

import json
import matplotlib.pyplot as plt

with open('input.json') as fp:
    json_data = json.load(fp)

for v in json_data['Viewers']:
    plt.plot(json_data['Dates'], v['Viewers'], label=v['Name'])
    
plt.legend()
plt.xticks(rotation = 20)

# to save as png, it should be done before plt.show()
plt.savefig('graph.png')
plt.show()

And here is the result

CodePudding user response：

Assuming original question is exact, meaning plot needs to show 'Dates' in Legend, and data should not include Total, only data for each user:

understanding the data:

  import pandas as pd 
  import plotly.express as px
  data = [...]
  dates = data['Dates']
  users = [x['Name'] for x in data['Viewers']]
  viewers = [x['Viewers'] for x in data['Viewers']][:-1]
  data_list = [(dates[index], users[index2], y) for index, x in enumerate(viewers) for index2, y in enumerate(x)]
  df = pd.DataFrame(data_list, columns =['Date', 'User', 'Viewers'])

plotting it:

  fig = px.bar(df, y='Viewers', x = 'User', color = 'Date', barmode="group")
  fig.show()

result: Plot with 'Dates' as legend

That being said, a time series is usually intelligible if timeline is on x axis. However, given the (interview/technical test?) requirements, I believe this is the correct answer.