let me start out by saying I have a basic understanding of python and excel. What I’m trying to do is take a CSV file with data points for voltages along side with date in mm/dd/yyyy in a separate column, and averaging the data points for each day into a chart. There’s roughly 10000 data points for the month each separated by 5 minute increments. I don’t know where to start. I also have to do this about 12 separate times so if I can just feed a program an excel file or data from notepad that would help me tremendously.
I’m a bit rusty with python but I can read it pretty well to figure out what does what. Any help would be greatly appreciated.
If I can do this all in excel then that would be amazing. But I would like to brush up on my coding. If someone can point me in the right direction, I’m using Jupyter notebook and excel for reference. I tried starting with just opening the file with open() but now I don’t even know where to begin. Should I install pandas? Selenium?
I tried grouping the data in excel manually by using group. I tried opening up files in Python and creating a list of values to go through and average them. I also looked into using pivot tables to group the data as well.
No coding needed for that. Excel pivot can solve it.
Bring your date column in a format excel understands as date.
Insert pivot from the range that contains your data
Drag the Date to lower left quadrant rows
Drag the voltage to lower right quadrant values
Change aggregation of values from sum to average
The result will be this table:
Excel seems not to enjoy the quantity of data i need to plot.
So i am considering, is it even possible with excel?
I have tried using pivot tables, but it splits the data into multiple parts, which i dont need.
I also tried Using a normal line plot, i get the error maximum 255 in a series.
Do i need to find alternative software for this plot?
Example of the data:
Long time open question, in the mean time i found a answer which i will share.
I went outside Excel and found a old outdated app called DatPlot
and it worked like a charm..
I think this problem, when solved by creating additional charts with offsets, is easy. I want to cut out the middle man and not use offsets (unless they are useful to the answer). I have data for daily cohorts and I know specific information about their behavior 1 day later, 2 days, 3 days ect.
Now it is rather easy to make a waterfall chart of day by day activity like so...
What I want to do is skip this step (directly above, the waterfall chart) in hopes of shrinking my current workbook by a substantial amount. You can imagine having simply 1 year of data across multiple channels measuring even 1 aspect of behaviors can account for a lot of data and pivot charts. Also, btw, I have the top chart as a pivot thus allowing this to be hands off when calculating what I am looking for.
What I seek - I look to further construct groups of days as other cohorts to examine (for example, say, 1/1 - 1/5) and see what their activity has been in a cumulative fashion since then. To be more specific, I want a table that will show cohort 1/1-1/5's activity in the date range 1/1-1/5 (11) and then their activity from 1/1-1/9 (24, an additional 13 "behavior points" summed).
So far, as I said, my current solution involves the "blue arrow" schematic where an additional table is constructed and I can sum on, essentially, rectangles build by using OFFSET on sell ranges with the MATCH function. I am stumped with how to go about this without the additional charts.
Thanks!
VBA would better for this, but use this formula in C30:
=IFERROR(SUM(SUMIF(OFFSET(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(C$28)),SEQUENCE($B30-$A30+1),IF(COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(C$28)))-SEQUENCE($B30-$A30+1,,0)>COLUMN($B$1),0-SEQUENCE($B30-$A30+1,,0),COLUMN($B$1)-COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(C$28)))+1),1,IF(COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(C$28)))-SEQUENCE($B30-$A30+1,,0)<=COLUMN($B$1),(C$29-C$28+1)-SEQUENCE($B30-$A30+1,,-(COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(C$28)))-COLUMN($B$1)-1)),C$29-C$28+1)),"<>")),0)
and this in D30:
=C30+SUM(SUMIF(OFFSET(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(D$28)),SEQUENCE($B30-$A30+1),IF(COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(D$28)))-SEQUENCE($B30-$A30+1,,0)>COLUMN($B$1),0-SEQUENCE($B30-$A30+1,,0),COLUMN($B$1)-COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(D$28)))+1),1,IF(COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(D$28)))-SEQUENCE($B30-$A30+1,,0)<=COLUMN($B$1),(D$29-D$28+1)-SEQUENCE($B30-$A30+1,,-(COLUMN(OFFSET($B$1,$A30-MIN($B$2:$B$10),MIN($B$2:$B$10)-$A30+DAY(D$28)))-COLUMN($B$1)-1)),D$29-D$28+1)),"<>"))
And copy both down.
If one does not have the dynamic Array formula SEQUENCE() then replace all the SEQUENCE($B30-$A30+1) and SEQUENCE($B30-$A30+1,,0) with ROW($ZZ$1:INDEX($ZZ:$ZZ,$B30-$A30+1)) and (ROW($ZZ$1:INDEX($ZZ:$ZZ,$B30-$A30+1))-1) Respectively, and use Ctrl-Shift-Enter instead of Enter when exiting edit mode.
I was able to collaborate on a solution. I am told that it will be highly inefficient at scale but it gets the job done. It ss less automation-friendly but can be formulated to capture data not currently present on a, say, a pivot table that you call to update later by extending the area that the formula works on.
Formula in in I31:
=SUM(IF(($C$1:$O$1+OFFSET($B$2,$G31-$B$2,0):OFFSET($B$2,$H31-$B$2,0))>=I$29,OFFSET($C$2,$G31-$B$2,0):OFFSET($O$2,$H31-$B$2,0)))-SUM(IF(($C$1:$O$1+OFFSET($B$2,$G31-$B$2,0):OFFSET($B$2,$H31-$B$2,0))>I$30,OFFSET($C$2,$G31-$B$2,0):OFFSET($O$2,$H31-$B$2,0)))
I have a report with 2 pie charts, giving sales data for product groups. Running a dataset from an OLAP cube.
I want to be able to select 2 different dates for each chart.
Example; Someone wants to compare November sales to the previous month. So one chart to display October, one to display November. Or they might want to see how Jan/Feb/Mar compare to Apr/May/Jun.
Can this be done from the same dataset, or do I need to point each chart to a separate dataset?
If the dataset returns the data for both months, you could add filters to each chart to show the appropriate one. However, it sounds like you'd want this to be more dynamic.
I would suggest creating two separate datasets, each with a month parameter. This will allow the user to select any combination of two months. The charts would simply point to their respective dataset. You shouldn't incur a noticeable performance hit either way.
I made a table with my earnings and cash disbursements. And I list the account balance.
Now I want to draw the account balance in a line chart. But the problem is that there isn't a value on every day. So the line between to entries is different because different time lies between two entries. Sometimes three days, sometimes 22 days, ...
How can I realize, that the line is constant after a value and the line only change on a day, with a new value?
Edit (from comment): I'm using Excel 2007
Like the others have said, to get the true time difference between your points to show up, you have to use a scatter plot instead of the Excel line graph.
To display the true state of your balance at in-between dates (i.e., no sloping lines), you'll need to do some transformation of your data. It amounts to creating two new columns of data with duplicate points for each balance: one at the date the account balance reached that level, and one at the date the balance left that level.
I suggest the following:
To populate your Balance column, in the first cell (I'm starting my table at E2) enter
=B2
where B2 is the address of the first balance in your original table. In the next cell down enter
=IF(E2<>E1,E2,OFFSET(B$2,COUNT(E$2:E2)/2,0))
Fill down this formula as far as you need to.
Then in the first cell of the Date column of your new table (I'm starting at D2) enter
=A2
where A2 is the first date in your original table. In the next cell down enter
=IF(E3=E2,OFFSET(A$2,COUNT(E$2:E3)/2,0),D2)
Fill down this formula. Then use this new table as your source data for the scatter plot. Here's a screen capture of a sample:
Are you using an XY-Scatter Plot. A line chart adds in the missing days automatically. Make sure your x-axis is formatted as date.
If you insist on using XY-Scatter Plot then you can put in days for each day, if that day doesn't have data then for the y-axis put in =NA().
I think the chart you need is a step line chart based on your description. The answer of #Excelll above has already pointed out a way to do this in Excel. But here I've another solution that could avoid manually adding some rows to your data. You could use the Funfun Excel add-in to create step line chart. Here is an example that I draw.
As you could see there is some code in the Excel. The Funfun Excel add-in allows you to use JavaScript code directly in Excel so that you could use powerful libraries like HighCharts.js or D3.js to draw charts that could be difficult in Excel alone. In this example, I used HighCharts.js. And it's quite easy to plot this chart since HighCharts.js iteself supports step line chart, so all you need to do is add the step option in your code like the code below.
series: [{
name: "Balance",
data: balance,
step: true,
showInLegend: false
}]
Another advantage of the chart that I showed you comparing to other solution is that you get real time interval in your x-axis. As you mentioned, you may not take record of your balance everyday, the x-axis in the example above shows irregular time interval.
The Funfun also has an online editor in which you could explore your JavaScript code and result. You could check the detail of how I made the example chart in the link below.
https://www.funfun.io/1/#/edit/5a4e478f1010eb73fe125cb2
Once you are satisfied with the result you achieved in the online editor, you could easily load the result into your Excel using the URL above. But of course, first you need to add the Funfun add-in into your Excel by Insert - Office Add-ins. Here are some screenshots showing how you load the example into you Excel.
Disclosure: I'm a developer of Funfun
Excel is really rubbish at charting this kind of thing. In my experience, the best way to chart it is to start with an X-Y plot rather than a line graph, and select the option where it joins the points with a line.
You may also need to add a new column showing "days since start", and use that as the X axis instead of the date - I can't remember whether Excel 2007 can deal with dates properly or not. Try it and see.
Finally, if you include any cells in your chart that are not yet populated (eg because you want to make the chart from the whole column even though you only have a few entries so far), you need to have the empty cells populated with "#N/A" otherwise it will plot them as 0.