Planning Analytics

Planning Analytics

Get AI-infused integrated business planning

 View Only
  • 1.  Implementation of demand forecasting solutions in Planning Analytics for utility sector

    Posted Mon February 01, 2021 10:37 AM
    Hello All,

    I have a requirement to Implement demand forecasting solutions in IBM Planning Analytics for the utility sector. I would be grateful if you could share any referenced document related to the same.

    ------------------------------
    Shruti Gharat
    ------------------------------

    #PlanningAnalyticswithWatson


  • 2.  RE: Implementation of demand forecasting solutions in Planning Analytics for utility sector

    Posted Mon February 01, 2021 12:46 PM
    Shruti:

    Are you looking for predictive demand forecasting or driver-based revenue forecasting, or potentially both? :-)

    You can find out more about Predictive Demand Forecasting using IBM Planning Analytics (PA) on QueBIT's website here (including a free white paper): https://quebit.com/demand-planning-for-enterprise-businesses/predictive-demand-planning-and-forecasting/ - our solution "Galileo" is basically a predictive solution integrated with PA.

    We have also helped a number of utility customers with revenue forecasting but unfortunately do not have any public write-ups. Reach out if you would like to discuss.

    Regards,
    AG

    ------------------------------
    Ann-Grete Tan
    Chief Strategy & Marketing Officer
    QueBIT Consulting

    Analyze. Plan. Achieve.
    https://quebit.com
    ------------------------------



  • 3.  RE: Implementation of demand forecasting solutions in Planning Analytics for utility sector

    Posted Tue February 02, 2021 03:15 PM
    Hello Ann-Grete Tan,

    Glad to receive your reply. As per your suggestions, I visited your website and took a glimpse at your solutions including the white paper. The solution is really great but unfortunately, the user wants everything in Planning Analytics without integrating it into any other solutions.

    Furthermore, I am looking for an end-to-end solution of Planning, Budgeting as well as Forecasting for the utility sector. I would be pleased if you assist me further on the same.

    ------------------------------
    Shruti Gharat
    ------------------------------



  • 4.  RE: Implementation of demand forecasting solutions in Planning Analytics for utility sector

    Posted Tue February 02, 2021 04:30 PM
    Hi Shruti

    I did something for electricity a while back. Are you working for the utility or are you a consultant/contractor? What sort of utility is it? Electricity can have some quite different drivers to water or gas. You are probably going to want to factor in the weather forecast for these. That generally means at least some external data. Phone / broadband is completely different again. 

    Regards

    Paul SImon

    ------------------------------
    Paul Simon
    ------------------------------



  • 5.  RE: Implementation of demand forecasting solutions in Planning Analytics for utility sector

    Posted Wed February 03, 2021 04:40 AM
    When I worked for Applix (the owners of TM1 before IBM bought the product) we built a Power Purchase Agreement (PPE) application for a major UK based energy provider.

    A PPE is a contract between different energy providers to trade energy to meet demand when the host has insufficient capacity to supply the demand, or the cost of starting up a tranche of a power station to meet demand is more expensive than buying it from another energy company.

    An energy provider will have multiple PPE contracts with multiple providers so the calculation logic is unique to each one and are probably the most complicated TM1 models I've seen from a Rules perspective.

    In the models we created, in addition to the 5 year daily time dimension there was also a half hour timeslot dimension to break down each day, these are known as Settlement Periods and demand would be split into these half hour timeslots too. There is also the ability to have 50 half hours in a day, 46 too as you have longer and shorter days for days when the clocks change for daylight saving time, of course these are on different dates each year.

    Days are often flagged as Winter or Summer days using Attributes and Settlement Periods are often flagged with Day and Night attributes and in this model most formulas had 4 IF statement e.g. If Winter and Day THEN, else IF Winter and Night THEN, else if Summer....

    As far as I'm aware, the model didn't forecast demand, but costed the cheapest way to meet it, either by supplying it themselves us buying it from another provider.

    As Paul mentioned earlier, forecasting demand is complex and requires multiple inputs e.g. Weather, is there a football game on (everyone will switch their kettle on to make a drink at half time) and I'm sure they had other software that did this sort of work for them, maybe something like SPSS instead.


    ------------------------------
    Mark White
    ------------------------------