Large survey - long loading time

Hi

We have developed a rather large survey.
It contains 1139 questions
(when data is exported to Google Sheet it contains 1597 columns)

The data collector is using a smartphone for collecting the survey data.
They open enketo for collecting data.

The issue for the data collector is that the empty form takes a “long” time to load (1-2 minutes - my last test was 1 minute 31 sec )

In some cases the data collector saves the form as draft. (Which takes only seconds - except the popup telling you the draft is being saved does not dissappear for a long time or the screen freeze for roughly a 1 min 30 sec.)

When they want to open the draft it takes again a very long time (my last test 1 minute 42 seconds).
I am testing this on a low-end small tablet. But the data collectors are saying it can take up to 5 minutes to load a draft.

The first issue is that the data collector think nothing is happening, because when you select to load the saved draft - there is no indication that the draft survey is loading. Please consider showing some kind of hour glass while loading, so it is clear that something is happening.

Is there some hints on how to speed up load time of the saved draft form?
the form contains

  • many calculations
  • a few (3) user uploaded images (max size is not set !)
  • 6 select from file questions (facility list (10.000 units) and administrative units of Nepal) which are stored in 5 different CSV files (See attached for file sizes)

I can try to optimise the survey changing one of each of the above issues, but if some issues are known to create loading problems, I would like to try that first :slight_smile:

This is a link to the .xslx file in my Google Drive open for comments 2023-01-28 SPARS Supervision tool - Google Sheets

Any suggestions are most welcome

@steen_andersen, generally, it could time out when you have a large survey (with many questions and complex constraints/expressions). The best possible option to handle this is to try to break the long survey project into multiple surveys (which does not distract/distort the objective of your survey).

Perhaps you may also review the design of your form and try to use repeats instead of duplicating the same question groups. This may also facilitate combined check rules, calculations and analysis.

1 Like

Thank you for your suggestions.

Splitting up the survey is aa good suggestion. This should make loading etc faster.
But (always a but) it would make it a bit harder to train the data collectors, to understand they need to open 2-3 surveys to conclude - and it is not a possibility, because we need some calculations done across all indicaators at the end of the survey.

It may be a possibility to change two indicators to be repeat groups. I will have to look into that.

For now I have asked the data collectors NOT to save as draft. I see no reason for them to save the survey as a draft before submitting it. They are not supposed to make changes to the data once they leave the facility!

1 Like

One question though - will it change a lot if I remove the “select_one_from_file” and convert them into simple cascading choices? The file method was choosen to have all facilities in the country available - by in reality we are only collecting data in 3 out of 7 provinces!

Please do full pretests, for the complete process, includind data export for analysis, before starting a survey .

1 Like

Dear Wroos
thank you for the valuable advise.

1 Like