Bulk import objects (csv files) into groups

Question

I'm looking for examples of software and best practices around bulk import to groups. To keep the project confidential let's say:

John is a private delivery man. He has 20 deliveries today that are split between 3 customers. His onboard computer automatically creates a .csv file every time he makes a delivery. At the end of the day he has 20 .csv files.

What I'd like to do with our software is bulk import all 20 of those files and group them into Jobs 1, 2, and 3 in the most automated way possible. The driver will still have to manually input some data such as road conditions etc. The user will need to have control over which objects belong in which groups.

Are there any best practices or examples of software that solve similar problems.

Answer 1

I'm guessing a lot here - if you find my assumptions incorrect, please edit the question to add more information.

This is a general answer, because I do not have enough information to provide a "design". My answer is more about the overall process, and I would like to encourage thinking about changes to the process which will allow the final interaction to be as easy as possible.

Assuming you have no control over the complete process, and can only influence the last step, where the delivery person classifies and updates data after import: You still have two general options, (1) provide all data in a tabular format and allow editing in the cells where corrections are allowed or required, and (2) present one delivery after the other (a form instead of a table) and let the user correct each one individually. Choice between (1) and (2) depends on number of required manual corrections, number of deliveries (and the ratio between these numbers), the device this shall run on (read: available space), the nature of changes (select from 3 choices, or 100 choices, or free text), and possibly more.

Assuming you also have control over the content of the imported files, make sure to pick up as much relevant information. I think the grouping should be done from the delivery orders automatically (look at the sender, or whatever), for example. Road conditions may be derived from a ratio between travel distance to travel time. If you think hard, you'll find ways to propose good defaults which must seldom be changed manually.

Assuming you can also influence the on-board process, get rid of the import step entirely: Let the driver create the final data when he comes back from the recipient's house, before he drives on to the next delivery. In that way, he must not remember all this until the evening, which will degrade data quality. Again, provide as many defaults as possible so the driver only has to nod instead of entering information.

Answer 2

I think a good/decent design can easily be achieved for this scenario.

1. Importing N number of .csv files can be as simple as importing all files within a user-provided folder.

2. Grouping N number of .csv files and grouping them into K number of Jobs is trivial. You can get fancy here, but for now a simple algorithm that distributes evenly among K groups should be fine.

3. Present the imported data as rows within a datagrid. Allow multi-selection and assign the selection as a Group (color-coding the groups would be a nice bonus). You could have a separate datagrid per Job (e.g. "Job 1 of 3").

4. Allow columns to be edited AND the option to apply a change to the entire group.

In short, the general rule of thumb should be keep-it-simple. Keep it simple, neat, and elegant. Then hopefully users will find it intuitive.