Sivaji

Within our Model & Predict tool we can set our training window and train and get the statistics of the trained model, however also having the feature to have testing period and providing statistics during testing would be nice to have. So, that users can see if they are able to generalize their and avoid over fitting.

I noticed the following 1. I created a current workstep link in the journal -- this works 2. I made some updates in the worksheet, and now if I want to replace that workstep link, if I try adding current workstep link to the already created workstep link, it doesn't work. Also no notification saying we cannot replace. I was aware that I can delete the existing one and create a new link. But, as an user, I prefer having a replacement capability like the traditional hyperlinks.

Currently I can able to create several files with same name. Although the IDs of workbooks are different, but it would be nice for SEEQ to not to consider same name files within the same folder as we do in our traditional folder structure. Also as previously suggested, it would be nice for SEEQ to pop-up asking for naming the Workbench analysis/Datalab project when we create, instead of creating a default name based on timestamp and then coming back and editing the name.

Hi, I was working on an external calculation with a sliding window, however, as per the documentation, the window is forward-looking. I'm interested in a backward-looking window instead of a forward one, I tried to input '-' before the window size in the formula, but it threw an error message. This will help to calculate signal values up to the latest samples available. Thanks,

Hi, I have an use case where I need to calculate aggregated value of every n-samples and replace the samples with the aggregated value in case the difference between the sample and aggregated value is greater than certain threshold.

When I place python scripts (.py extenstion) in the DataLab project folder and try to run those scripts using the DataLab project terminal it throws an error saying to spy.login first. However I don't require spy login for running jupyter notebook files (.ipynb extention). I run test cases using python files, it would be nice feature to be able to run these test python scripts from the DataLab without doing a spy.login.

From the documentation of external calculations it looks like we don't have an option to return multiple output signals, this feature can be useful if I want to return the intermediate calculations that I used for the final output. The wrokaround as of now can be having multiple external calculation scripts for each of these intermediate steps and then use these signals as an input for the final output external calculation script.

Hi, I was wondering if we can do all the SEEQ calculations within DataLab itself without pushing and pulling to SEEQ? For example if I have input signals as a dataframe with timestamped index and I wanted to apply some-kind of SEEQ formulas for example like a advanced value search, as of now I do following steps 1. Push the input signal to SEEQ 2. Apply formula from DataLab on the input signal and push the formula to SEEQ 3. Pull back the results of the calculated formula from SEEQ. I wanted minimize or avoid SEEQ interactions and calculate the results directly in the DataLab. For basic calculations I can use equivalent python/pandas functions, however for SEEQ specific calculations I need to follow the above procedures. Is this feature available now or in the roadmap? Thanks, Sivaji

Thanks, Thorsten. I was looking for something like the below search and look at the signal trend.

I was wondering if we can do a Data search a signal (formula/condition) by its ID? Use-case: for validating IDs used in the DataLab project created by others.

As part of formatting organizer reports, I think having an option to hide or white color cell borders might be an useful feature. This way users can import worksheets/images in the cell boxes for alignment and can hide the cell borders to make the report look better when exported to PDF. This feature is not available as of version 51. As of now, only Dashed/highlighting and thick options only available.

I was wondering if we can have a feature to create a new/duplicated worksheet next to the current selected one instead of creating it at the end of the workbench? Some cases when we have multiple number of worksheets within a workbench analysis, it's hard to move worksheet from the end position to the required position. I was looking for a feature like a traditional power points/google slides where the newly created slide will be next to the current selected slide.

Thanks Joe, yeah forecastLinear works for linear forecasting. I'm looking for an option in "Model & Predict", for a "Time Series Forecast" which can be similar to the "Prediction" feature, however the only change I'm looking for is there shouldn't any option for "Input Signals", this feature should take the time series as the input signal by default (considering the training window). I am aware that we can create a signal using timesince() function and use this as my "Input Signals" and use the traditional "Prediction" feature. However SEEQ mostly deals with the time-series data, I think having this feature could give benefit to the customers who wants to do some kind of Time Series Forecasting of their signals. Customers might be interested in predicting sinusoidal/seasonal trends also along with linear trends. Again this is a suggestion.

I understand in the current version we need to create a time signal using timesince() function, and use the predict function to train the signal that we need over the time signal created. However I was wondering if we can time-series forecast function which internally creates the time signal based on the start datetime of the training window and trains the model.

When I tried pulling it the result showed "Not a Signal, Condition or Scalar - skipping". Alternatively I can create a signal from condition and can pull the data. However I have already calculated many scorecard metrics formulae, I was wondering if there is any workaround to pull them using DataLab?