Protobi

← All help posts

View post

### Clients request
**Q.How many layers of nesting can you have in a banner?**
**A.** In Protobi, you can have up to two levels of nesting in a crosstab banner. Here are the key details regarding banner nesting:
**Two-Level Limit: **You can add one main banner and then a second "+" button will appear, allowing a second nested banner (e.g., Specialty > Region).
**Method to Add Nested Banners:** Drag the first banner variable to the crosstab area. To add a second, hold down the shift key while dragging the second banner over the existing banner.
**Reordering:** You can double-click on either banner to swap the nested order.Multiple Columns: If a second banner is added, the columns are sorted by the first level banner, and then the second.For more complex analysis, you can also use SuperTables to create more extensive crosstabs, though the standard interactive banner generally supports two levels
Protobi is gives a concise view so we are limiting it here to two levels but in Protobi sql we can create as many nesting as you want.

**Q: If two people are working in the same project at the same time, will they automatically see each other's changes as they save?**
**A:** No. Changes do not automatically appear for all users who have a project open. When a user saves changes, the bubble in the upper right corner turns yellow this is a signal that everyone with the project open should reload their browser tab to see the updates.
It's important to save your own work *before* reloading, otherwise any unsaved changes will be lost (reloading fetches the last saved version from the server). For this reason, **simultaneous editing is not recommended**  if multiple people edit and save around the same time without consistently reloading between saves, they risk overwriting each other's work.
Protobi does, however, keep a **local history** of your changes. You can access it by pressing **Shift+Z** or selecting "Local history" from the toolbar context menu. This lets you restore your project to any point during your session even if you accidentally reloaded without saving. If prior saves have already been committed to the server, those can be recovered from the **"History"** section on the project settings page.
**Q: If I reload to see someone else's changes but hadn't saved my own edits, do I lose my work?**
**A:** Yes. Reloading the browser tab without saving first will fetch the last saved version from the server, and any unsaved local changes will be lost. Always click **Save** before reloading to commit your work.
**Q: If I save before reloading, do I save my work but overwrite the other person's changes?**
**A:** Yes. If a colleague has already saved their changes and you save without first reloading to pull in their latest version, you will overwrite their work. The correct workflow is: **reload first** to get the latest saved version from the server, then make your edits, then save.
* The number showing in the yellow circle does that mean someone else is actively in the file?                   
* Can you specify which waves get rolled up when combining banners?                                                
* If I don't need a particular banner, can I just remove it from the Super Table export?                           
* This export ran data for the whole survey with this banner is that the idea?                                   
* Any data manipulation we want to do should that be done before creating the banner and exporting?              
* Is there a way to export without stat testing?                                                                  
* The export opens in Excel is there also an online view?                                                       
* Is there a tips and tricks guide or knowledge base so I don't have to ask every time I want to do something basic? 
* Is a global crosstab like starting a banner without formally building one?                                      
* Can you filter quickly to just one specialty or a combination of waves?                                         
* Can we set up the project with an interim or partial dataset before the final data is ready?                    
* If we load partial data midfield and then get the final dataset, can we just replace it without rebuilding everything? 
* If we want to keep the partial data on record and not overwrite it, is that possible?    
* How do I combine ratings from current users and non-users into a single variable for each treatment?

**Q** How do I combine ratings from current users and non-users into a single variable per treatment? one is 2dgrid,one is 1dgrid,one is atomic
**A** When you have two separate questions one asking current users about their satisfaction (C12) and another asking
   non-users about their perception (D5) you can create a new combined variable for each treatment that merges both
  groups into a single 7-point scale.                                                                                 
                  
  The logic for each new variable is:  
  - If the respondent answered D5 for that treatment → use their D5 response as-is (including "don't know" codes)     
  - If D5 is blank but they are a current user of that treatment (e.g. S11 = 3) → use their C12 response for that
  attribute                                                                                                           
  - Otherwise → leave blank                                                                                                         
  This gives you one variable per treatment that covers anyone who is either knowledgeable about or has direct        
  experience with that treatment.                                                                                                               
  Things to keep in mind:                                                                                             
  - Do not override the original D5 or C12 variables create a new variable (e.g. dx_[product]_[attribute])
  - Preserve any special codes such as "don't know" (98) from D5 in the new variable                                  
  - If you need to export the combined variable to SPSS for driver analysis, build the logic in the data process
  rather than pre-calculate    
	
**Q** D10 is a 7 point scale question where we want to show the net from a question s11_a with those who say they are **not aware of the treatment** How to implement this?
**A:**
First we turned on the NA from D10 to do it permanently in any element we need to go to the context menu and from edit properties we need to we should set showMissing to yes, and now we go back to s11 and clicked on 7% which is not aware of the treatment and we go back to D10 and when we removed the filter by pressing esc and saw everyone answered s11
The logic is:
If you answered s11_a, if you put the value 1, I'm going to reassign you a value at D10_a I'm not going to give you a value of NA. I'm going to give you a value  And that's going to be our don't know.
Then we can write a 
example way:-
 if (row["s11_a"] == 1) row.d10_a = 99 //don't know
       if (row["s11_b"] == 1) row.d10_b = 99 //don't know
       if (row["s11_c"] == 1) row.d10_c = 99 //don't know
       if (row["s11_d"] == 1) row.d10_d = 99 //don't know
       if (row["s11_e"] == 1) row.d10_e = 99 //don't know
       if (row["s11_f"] == 1) row.d10_f = 99 //don't know
Now we should go to the format and name 99 as dont know.
**Q:** How to show the row tests comparing treatments A, B, C, D, E, F.
**A :** We need to clone the chart and select paired table from chart type.
Steps:
1. Clone the chart
2. Go to **Advanced → Clone Simple**
3. Open cloned chart
4. Change Chart type → **Paired Table**
This enables:
* Row comparisons across treatments
* Statistical testing between columns

Publishing

Date

Status

Slug edit

Content

Thumbnail

Categories Manage

Tools

AI Database

Publishing

Content

Tools

Danger zone