Fact Table Data Structure for non-business Time Stamp Analysis help

Hi all,

I am new to PBI, but have been using Excel for many many years. I recently become interested in analyzing some time stamp data in PBI for its cloud based reporting rather than in Excel, and I'm in need of some advice. I have gone through many tutorials and I believe I understand how to use a star schema for the tables, and I believe that I can work out my "Dimension" tables quite easily. Where I need advice is the structuring of the "Fact" table.

Background: I analyze data that is generated from behaviors recorded on video, applying codes and timestamps to the video, and exporting these time stamps. The difference in time between some of the behavior codes is the meat of the data that is used.

Here is an example of the exported data:

I have queried the data into PBI, and done some transformations, shown in the following M code:

let
    Source = Csv.Document(File.Contents("FILENAME.csv"),[Delimiter=",", Columns=4, Encoding=1252, QuoteStyle=QuoteStyle.None]),
    #"Promoted Headers" = Table.PromoteHeaders(Source, [PromoteAllScalars=true]),
    #"Changed Type" = Table.TransformColumnTypes(#"Promoted Headers",{{"Observation id", type text}, {"Behavior", type text}, {"Behavioral category", type text}, {"tStamp (s)", type number}}),
    #"Split Column by Delimiter" = Table.SplitColumn(#"Changed Type", "Observation id", Splitter.SplitTextByDelimiter(" ", QuoteStyle.Csv), {"Subject", "Event"}),
    #"Changed Type1" = Table.TransformColumnTypes(#"Split Column by Delimiter",{{"Subject", type text}, {"Event", type text}}),
    #"Reordered Columns" = Table.ReorderColumns(#"Removed Columns",{"Event", "Subject", "Behavior", "Behavioral category", "tStamp (s)"}),
    #"Added Index" = Table.AddIndexColumn(#"Changed Type1", "Index", 1, 1, Int64.Type),
    #"Added Custom" = Table.AddColumn(#"Added Index", "NextObs", each if [Behavior]="ST" then 1 else 0),
    #"Added Custom1" = Table.AddColumn(#"Added Custom", "ObsID", each List.Sum(List.FirstN(#"Added Custom"[NextObs],[Index]))),
    #"Removed Columns" = Table.RemoveColumns(#"Added Custom1",{"NextObs"}),
    #"Reordered Columns1" = Table.ReorderColumns(#"Removed Columns",{"Index", "Subject", "Event", "Behavior", "Behavioral category", "tStamp (s)", "ObsID"})
in
    #"Reordered Columns1"

you can see that this code adds an index, and some colums to add an observation ID that increments up every time there is an ST behavior code. This is simply because each observation starts with the ST code. in order to look at the specific quantities that I want to analyze, I need to take the difference in time between the tmiestamps of specific behavior codes.

Here is an example of what I mean:

Previously I was doing this analysis in Excel, and basically parsing the vertical data from each observation into its own row using a macro like this:

This leads me my questions:

• In order to be able to analyze the quantities for each event, subject, or even observation, do I have to build the same type/structure table of the data as I have done in excel? Or should I be able to keep the data structured as I have it in my query?
• If I shouldn't structure a fact table like I have done, what is the most efficient way of doing the quantity calculations?
• Is there a best practice for structuring the data for an analysis like this in PBI?
• are there any specific tutorials I should be looking at? I've done/watched a bunch of standard tutorials, but I am having trouble getting past this simple step!

If necessary I can supply example data and pbix files as well.

thanks in advance!