cancel
Showing results for 
Search instead for 
Did you mean: 

End User Examples of Prediction Model

pawlowski6132
Blue LED

So, I've researched this on and off for a few months while we slowly begin using DatRobot and can't find the answer I'm looking for. In fact, this issue has come to a head now and we are stuck.

 

Our initial Use Case problem to solve was shipments out of a warehouse that had quality issues (incomplete, damaged, etc.). We (central, not at the warehouse) developed a model to predict the at-risk shipments such that an operator at the site can take actions to remediate.

 

So now we have a model in Datrobot but, now what??? How can a shipment supervisor upload scheduled shipments into the model and get a short list of shipments that are at-risk?

 

I've read many of the Deployment articles in the knowledge base and watched some of the videos in the training section but, they only seem to go as far as creating something just short of the end user experience.

 

Thank you in advance

Joe, Detroit

3 Replies
jacob_a
Image Sensor

Looks like Datarobot just launched changes to pathfinder which has bunch of use cases. Maybe you can find something in there to help you. The 7.2 release notes mentions it https://community.datarobot.com/t5/blog/what-s-new-in-datarobot-release-7-2/ba-p/12469/jump-to/first.... I've gone through it quickly - it may have sometthing for you too https://pathfinder.datarobot.com/apps?industry=manufacturing 

 

Best of Luck

pawlowski6132
Blue LED

Hi, thanx for taking the time to reply. Good information but, really my question is more generic, i just gave a specific example. Basic question is: in summary, how do we operationalize or "deploy" our models to be used by real people in the business for whom we are developing these models? Is it not possible? Maybe Datarobot is not robus enough to deploy models to end users? thanx

0 Kudos
Lukas
Data Scientist
Data Scientist

Hi @pawlowski6132,

 

Thank you for your question. The best way of deploying a model will always depend on the specific use case and infrastructure. In your case I see two main options:

 

  1. Make use of our AI Apps. This will allow you to create an application, that can be used by end users. They can upload a csv file directly into.a visual front-end and analyse the results/download the predictions.
  2. Use the database integration to schedule scoring of new data. If the new shipments to be scored sit in a data warehouse, you can use the scheduled batch prediction option to always grab the latest data and push the resulting predictions back to a place that is accessible by your end users

Other options are available - someone from your dedicated account team will reach out to you to discuss.

 

I hope that helps.

 

Cheers,

Lukas