1

I have a situation where users are capable of creating their own searches on extensive data sets, using a query-like user interface. Our users come from a strong Excel background and are used to working with dozens or more columns in those.

Our search system provides a full screen interface with infinite scrolling and fixed headers, as sometimes these searches can return millions of records. We tested limiting the number of columns on a search to 12, but users were vocal about their desire to have more.

We get asked to add horizontal scrolling, but this works poorly with the existing system of fixed headers and infinite scrolling, in addition to generally making for a poor user experience.

Because the users select the columns themselves and can write custom columns using mathematical functions or aggregates on previous columns, it is difficult to predict the final appearance or content of the searches and not likely possible to condense multiple columns into a single cell.

Are there any patterns or shortcuts to provide what the customer is asking for without compromising experience or functionality?

  • Is it possible to classify column headers and use drop-down on them? – Mahm00d Aug 18 at 10:13
  • We can classify column headers by data type, e.g. decimal, integer, boolean, date etc. – Marc L. Aug 19 at 12:48
0

I think it's a good idea to categorize all the possible columns into a fixed number of categories and use drop-downs on headers:

enter image description here

Of course, it has a cognitive load on users to understand the categories and what each of them include.

0

You can definitely just make it similar to Excel, especially given that your end user is most familiar with Excel.

If you want to remove horizontal scrolling, then you can just hide or collapse the columns to ensure that all of them fit into the screen width. I will still suggest "make it exactly like excel".

  • Why do the users need this application if the have Excel? What is the value that this application brings? Is it a high enough value to compete with Excel? – Jennifer Aug 27 at 3:12
  • @Jennifer Normally excel data is very difficult to manipulate if you want to generate reports, or run data mining. Easiest option is always to upload excel to a relational database (think SQL server, OracleDB). Application like this can be directly connected to database, so that there is no manual upload needed. It also means you can have baked in reporting etc. At least that is the general thought behind it. Don't know about the specific requirement of the OP's application. – jitendragarg Aug 27 at 5:15
  • @Jennifer Like jitendragarg has said, the search component is linked to a relational database and other complex data types and is used to power reporting among other features – Marc L. Aug 29 at 10:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.