0

I'm working on a django web app which visualizes data stored in a SQLite in the browser using bokeh.

Since I want to display the data as a stacked area plot I concluded it might be helpful to use the pandas library to save the data from my django model to a dataframe before plotting.

My Django models.py includes the following model:

class Activity(models.Model):

    def __str__(self):
        return self.title

    title = models.CharField(max_length=50)
    sport = models.ForeignKey(Sport, on_delete=models.CASCADE)
    date = models.DateField(blank=False)
    duration = models.FloatField()
    distance = models.FloatField(blank=True, null=True)
    description = models.CharField(max_length=50, blank=True, null=True)
    trace_file = models.ForeignKey(TraceFiles, on_delete=models.CASCADE, blank=True, null=True)

Some example data in the SQLite db could look similar to:

| id |   title   |    date    | duration | sport_id | trace_file_id | description | distance |
|--------------------------------------------------------------------------------------------|
| 1  | biking01  | 2019-01-01 | 183      | 1        | 16            | NULL        | 142      |
| 2  | running17 | 2019-01-01 | 45       | 2        | 2             | NULL        | 14       |
| 3  | biking02  | 2019-01-01 | 67       | 1        | 18            | NULL        | 45       |

What would be the best way to read from the model and convert the data to a pandas dataframe? For example I have similar code in my views.py:

from .models import Activity
import pandas as pd

activities = Activity.objects.all().order_by("-date")
...

Similar to the dataframe in this example I'd like to get a dataframe which looks like the following, so I can plot a bokeh area plot.

   sport_id  duration  date
0  1         183       2019-01-02
1  1         67        2019-05-06
2  2         45        2019-03-04

Any help is appreciated.

bigreddot
  • 33,642
  • 5
  • 69
  • 122
Fabs
  • 161
  • 1
  • 9

1 Answers1

1

If you are trying to just simply export the data to visualize outside of the django app then check out ImportExportModelAdmin https://django-import-export.readthedocs.io/en/latest/api_admin.html

you can do a for loop to a list from each item and then set the list to a new df column 

import pandas as pd
df = pd.DataFrame()
game = []

for i in range(0,len(yourquery)):
    game.append(yourquery[i].game_id)
df['game_id'] = game

I haven't used bokeh before but I have used plotly.py and from experience the charts will load very slowly. Consider using something like plotly.js which can render similar charts but from json data you pass from the backend. A current project I am working on, which may be similar to yours, uses plotly.js site

iampotential
  • 121
  • 9