Word Frequency Analysis in Power BI (and R)

Sdílet
Vložit
  • čas přidán 12. 07. 2024
  • We analyse respondents’ comments of a recent survey and a Victorian horror novel. We use Power BI’s Query Editor to split the text into words. One stage of the process is to remove punctuation - this requires several steps in M but we add an R step to do this elegantly and simply.
    The datasets and course materials used in the examples are at github.com/MarkWilcock/lbag-o...
    Details and signup for events at www.meetup.com/London-Busines....
    Music: bensound.com

Komentáře • 27

  • @GTEM19889
    @GTEM19889 Před 2 lety +10

    Struggling to follow the step where you "split into words"... you mention you will come back to it but watched the whole video and I cannot work out how you did that from this video.

    • @vle987
      @vle987 Před 2 lety +5

      Under "Split Column" option there is an advanced option to split into rows using a delimiter

  • @xPanshishux
    @xPanshishux Před 3 lety +3

    Excelent ! I've learnt a lot with this video ! Thank you so much, greetings from Chile.

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety +2

      Glad it was helpful!

    • @raddleriddle7752
      @raddleriddle7752 Před 3 lety +1

      Hi great vid! But how did you get from text all in one cell to it being one word per cell? It’s like there is a step missing. Please help as this is what I really need

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety +2

      Hi James, @@raddleriddle7752 I used the Split Column operation in the Power BI Query Editor. Each video has a link to an example Power BI Desktop file that you can download and see all the steps

    • @stryder1587
      @stryder1587 Před 3 lety +1

      @@LondonBusinessAnalyticsGroup After splitting by delimited space, I have many columns, how do you combine them all into 1 column?

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety +2

      @@stryder1587 use the Unpivot operator.

  • @four321zero
    @four321zero Před 3 lety +2

    This is exactly what I wanted. Best video on the topic. Easiest to follow. Only thing I wish was that you could explain what a 'measure' is and what does adding the one you did, mean. Anyway, i blindly followed instruction and got my job done :D

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety +1

      Hi four321zero, A measure is just a fancy name for a number that we add up - such as the count of words. Glad you enjoyed the video. - Mark

  • @franciscorossier5239
    @franciscorossier5239 Před 3 lety +1

    Thanks for this video, it was a great help, and the documentation is perfect, downloading the example file, I learned a lot. The step using the scrip in R is still a bit of a mistery to me, but I guess I'll figure it out. Do I need to download/copy paste that script into R from some place before using it? Thanks and greetings from Chile.

  • @jasonq4254
    @jasonq4254 Před 3 lety

    Very excellent video, thank you! I got one question for you, how do you handle with plural form words when calculate word frequency? Maybe I miss something in your process?

  • @maleitne
    @maleitne Před rokem

    Very nice. Is there also a possibility to keep the survey IDs zu filter the surve by the word?

  • @sidh1973
    @sidh1973 Před 4 měsíci

    Excellent video, I learnt a lot! Thank you! What would be the advantage of building a WordCloud visual in PowerBI as against getting the WordCloud survey analysis from Microsoft Forms survey besides the flexibility of filtering?

  • @misslyceum
    @misslyceum Před 3 lety +1

    Thanks for a great video! Is there anyway to change plural words to singular, such as "servants" and "servant" are counted as one?

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety +1

      HI Tra, You could always remove the final s! You can do this in a R step using a R library like tidytext and a technique called stemming. - Mark

    • @misslyceum
      @misslyceum Před 3 lety

      @@LondonBusinessAnalyticsGroup thank you for your reply. Really helpful :)

  • @efficiencydna897
    @efficiencydna897 Před 3 lety +1

    thanks, great video. “split the text into words”,if my sentences are chinese such as “我今天吃了很多饭”,if there any method to do this?

  • @theway19
    @theway19 Před 3 lety

    How did you create word table? is it done manually? I have like 12000 rows in excel with comments, how am I going to create a word list out of that, please suggest.

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety

      HI theway19, I used R script in a Query Editor script to split sentences into words, then used the Query Editor Group By to count words. It's all in the video. - Mark

    • @LadyCrimson10
      @LadyCrimson10 Před 3 lety

      @@LondonBusinessAnalyticsGroup I watched the video but don't actually see the exact script that was used to split the sentence into words and distribute them in rows and not across columns.

    • @LondonBusinessAnalyticsGroup
      @LondonBusinessAnalyticsGroup  Před 3 lety +5

      Hi @@LadyCrimson10 , I used a "split column by delimiter" step. There is an advanced option to split into rows. Expand the advanced options section to see it.

    • @markcarmicunningham
      @markcarmicunningham Před 2 lety +1

      @@LondonBusinessAnalyticsGroup - "There is an advanced option to split into rows"... a nugget of pure gold

  • @vipul4raf
    @vipul4raf Před měsícem

    How did you created Stop_word table you have not mentioned. I think you have skiped some timline.