Assertion Error when producing ScatterText Visualisation

Question

I'm new to scattertext and have written the code which should produce an interactive html visualisation.

import spacy
import pandas as pd
import scattertext as st


twitterData = pd.read_csv("stock_data.csv")
twitterData.dtypes

nlp = spacy.load("en_core_web_sm")
corpus = st.CorpusFromPandas(
    twitterData, category_col="Sentiment", text_col="Text", nlp=nlp
).build()

sent = st.produce_scattertext_explorer(
    corpus,
    category="1",
    category_name="Positive",
    not_category_name="Negative",
    width_in_pixels=1000,
)

open("StockMarketSentiment.html", "wb").write(html.encode("utf-8"))

However, the code that I have written, following a template online, throws an assertion error and as I'm a newbie at software development i'm struggling to understand where Im going wrong.

Traceback (most recent call last):
  File "/Users/lukeashton/PycharmProjects/Project/venv/Visualiser.py", line 15, in <module>
    sent = st.produce_scattertext_explorer(corpus,
  File "/Users/lukeashton/PycharmProjects/Project/venv/lib/python3.8/site-packages/scattertext/__init__.py", line 594, in produce_scattertext_explorer
    scatter_chart_data = scatter_chart_explorer.to_dict(
  File "/Users/lukeashton/PycharmProjects/Project/venv/lib/python3.8/site-packages/scattertext/ScatterChartExplorer.py", line 115, in to_dict
    json_data = ScatterChart.to_dict(self,
  File "/Users/lukeashton/PycharmProjects/Project/venv/lib/python3.8/site-packages/scattertext/ScatterChart.py", line 276, in to_dict
    assert category in all_categories
AssertionError
Process finished with exit code 1

Appreciate it may be hard to offer advice with such limited info but the code & error details are below if anyone can spot anything!

Presumably `category="1"` is the problem. It seems it's expecting a different value other than `"1"` there. (Perhaps `"Sentiment"`?) — 0x5453, Jul 06 '21 at 17:51
@0x5453 hmm I see what you mean as per the error message, "Sentiment" throws the same error message however :/ I will try a few other possible values for that variable. — Luke Ashton, Jul 07 '21 at 08:22

score 2 · Accepted Answer · answered Jul 09 '21 at 21:44

2

Make sure at least in of the values in the sentiment column of your data frame is the exact string "1".

answered Jul 09 '21 at 21:44

Jason Kessler

116
3

The values in the column were int's so I converted them all to strings and voila! Thanks a lot for the pointer. – Luke Ashton Jul 13 '21 at 16:56

Dan Taninecz Miller · Answer 2 · 2022-07-06T20:09:12.863

Slightly different use case from the package, but I solved it this way:

html = st.produce_scattertext_explorer(corpus,
                                   category='dt',
                                   category_name='the thing you want to compare',
                                   not_category_name='other data',
                                   width_in_pixels=1000,
                                   height_in_pixels=500,
                                   minimum_term_frequency=2,
                                   metadata=df['parse'])

Where category='dt' is a dummy variable made into a col for whatever feature I want to pivot on in the df. I haven't had luck using an actual value from within a col in this way. The value here has to be a col in its own right to function without an assertion error (for me).

Assertion Error when producing ScatterText Visualisation

2 Answers2