为什么 plotly express 在时间线上抛出数据点？

Question

I have a simple dataframe with two columns. A sample is shown below, the data is available here .

   year-week  users
0    2018-22      2
1    2018-23      3
2    2018-24      4
3    2018-25      3
4    2018-26      5
..       ...    ...
69   2020-03    232
70   2020-04    226
71   2020-05    214
72   2020-06    203
73   2020-07    119

[74 rows x 2 columns]

When I try to plot those two columns with Plotly Express, it omits the data until 2019-30 .

import pandas
import plotly.express as px

df = pd.read_csv("https://pastebin.com/raw/x164p1Zp")
fig = px.line(df, x="year-week", y="users")
fig.update_layout(xaxis=dict(tickformat="%Y-%W"))
fig.show()

If I plot the same data with matplotlib, the data is shown:

import pandas
import matplotlib

df = pd.read_csv("https://pastebin.com/raw/x164p1Zp")
df.plot.line(x="year-week", y="users");

I cannot understand why two plotting libraries show the same data in a wildly different way.

How can I plot all the data points in Plotly Express to get a plot similar to what matplotlib shows?

Answer 1

Plotly does not recognize your x axis as being a date. You need to convert it explicitly to a datetime format.

Solution:

df = pd.read_csv("https://pastebin.com/raw/x164p1Zp")
# convert column to datetime, weekday needed for conversion to work
df["year-week"] = pd.to_datetime(df["year-week"] + '-0', format="%Y-%W-%w")
fig = px.line(df, x="year-week", y="users")
fig.update_layout(xaxis=dict(tickformat="%Y-%W"))
fig.show()