Redshift Serverless success, new documentation articles and ML

Hello everyone!
Another update from the DAP world, where we keep you posted about the latest development. With an increased number of users on the platform, as well as several new tables, the amount of support requests increases as well. This is good news. We strive to solve critical bugs asap, and we create tasks for future sprints for the less urgent ones. In order to strike a balance answering suggestions and bugs in a reasonable timeframe without disturbing the work already planned in the sprint, we have reduced the sprints to 2 weeks. "Non-urgent" issues reported to support are now typically handled in the "next sprint", but are of course subject to prioritisation, and some requests/suggestions may take longer, or be placed in the backlog.

Redshift Serverless:
Since the last update, we have successfully migrated to Redshift Serverless, which is a big technical upgrade to the platform. The most notable change for users is that we now scale limitlessly, and you should experience faster responses to your queries, independent of simultaneous traffic from other users. We can now also facilitate integrations with QuickSight. If that's something your organisation is interested in - let us know

Increased frequency of stream data ingestion

Stream data is now updated hourly! (Changed from 24h)

Machine Learning models: Parcel Volume Prediction POC

Now that we have high quality data on the platform, it becomes possible to do powerful analyses and predictions. We have developed a model for Parcel Volume Prediction as a POC for this kind of work, which provides quite good results already at the MVP stage. We have also set up the infrastructure to automatically retrain models at flexible intervals, to ensure the models get retrained and updated as new data comes in and the models become less accurate. If this is something that would be valuable to your organisation, or you want to learn more, please reach out. We would love to cooperate with you to find ways to utilise machine learning models and the surrounding pipeline infrastructure to solve the most pressing problems in your organisation.

Data model:
It is inspiring to see so much usage on the new data we have released! Salarydraft is one table we can see has a lot of value for you, as well as the retail data. We are excited to see increased traffic, and are grateful for the feedback you all provide, helping us iron out smaller and larger issues continuously, and providing suggestions for how to make the data model better and more valuable for everyone.

The columns previously known as route_number have been renamed to route_nr. In the past, having both column names would occasionally cause confusion. To ensure naming consistency, we have recently made updates to the affected tables.
Tables affected:

• Routeproductinfo

• Route

• f_plannedwork

• f_absencePeriods

A new table, DistributionPoint, has been introduced to the data model. This table contains delivery points that are associated with their respective modules. Its purpose is to serve as a link that enables the retrieval of all addresses and delivery points belonging to a route, regardless of whether they currently have active deliveries or not.

Freshdesk documentation

With the transition to Redshift Serverless, the user facing documentation has been updated, and moved to the DI Support site together with documentation from all other DI products. You can find it under Solutions -> Analytics -> DAP

Query sharing

Up until now you have been able to find SQL templates in the Web application. However, it is now possible for us to share SQL templates with you directly in the Redshift interface. We will move the templates to Redshift within the next few weeks, and publish a handful of new ones. We hope this will make it easier to get started finding value in your data, and act as an inspiration for further investigations. Some templates will be direct replica of popular analytical RG reports, intended to help you transition from RG reports to DAP much faster. The library of templates will be growing, and new additions will be documented in Freshdesk (see above), so make sure to check back regularly.

That's it for now, Happy DAPing!

Eirik Lyngvi
Head of Analytics