# About the Cold Start Energy Forecasting category
**URL:** https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556
**Category:** Cold Start Energy Forecasting
**Created:** [August 24, 2018, 10:43pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556 "2018-08-24T22:43:58Z")
**Posts on this page:** 16
**Page:** 1
### Author:  [@bull](https://community.drivendata.org/u/bull)
#### Post date: [August 24, 2018, 10:43pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/1 "2018-08-24T22:43:58Z")
This category is for posts about the [Power Laws: Cold Start Energy Forecasting](https://www.drivendata.org/competitions/55/schneider-cold-start/) competition. Please keep all posts here specific to the competition.
---
### Author:  [@ironbar](https://community.drivendata.org/u/ironbar)
#### Post date: [August 31, 2018, 1:21pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/2 "2018-08-31T13:21:48Z")
Hi
How is the winner selected? The best score on private leaderboard? Or do we have to select one submission?
Thanks
---
### Author:  [@vamsy19](https://community.drivendata.org/u/vamsy19)
#### Post date: [August 31, 2018, 8:32pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/3 "2018-08-31T20:32:28Z")
Hi
In the consumption\_train data, what does temperature column exactly mean?
And what does base\_temperature low or high for each series in the meta data exactly mean?
Thanks
---
### Author:  [@bull](https://community.drivendata.org/u/bull)
#### Post date: [August 31, 2018, 9:23pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/4 "2018-08-31T21:23:36Z")
@ironbar Winner is the best score on the private leaderboard.
@vamsy19 The `temperature` column is the outdoor temperature measured at the nearest weather station. The `base_temperature` is whether or not the setting for the building’s system is generally low (cooler) or high (warmer) that a certain threshold.
---
### Author:  [@sangxia](https://community.drivendata.org/u/sangxia)
#### Post date: [September 1, 2018, 6:00pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/5 "2018-09-01T18:00:16Z")
What happens to the evaluation metric if the ground truth is 0 (seems to exist in training set)?
---
### Author:  [@ashutoshshukla2894](https://community.drivendata.org/u/ashutoshshukla2894)
#### Post date: [September 3, 2018, 3:28pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/7 "2018-09-03T15:28:38Z")
We just have to predict that consumption is daily, hourly or weekly ? Is that correct ?
---
### Author:  [@bull](https://community.drivendata.org/u/bull)
#### Post date: [September 4, 2018, 11:55am UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/8 "2018-09-04T11:55:08Z")
In this case we would offset from zero by a very small epsilon.
---
### Author:  [@bull](https://community.drivendata.org/u/bull)
#### Post date: [September 4, 2018, 11:55am UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/9 "2018-09-04T11:55:53Z")
The submission format indicates what predictions we expect. For those time periods, you must predict the amount of consumption.
---
### Author:  [@sangxia](https://community.drivendata.org/u/sangxia)
#### Post date: [September 4, 2018, 3:22pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/10 "2018-09-04T15:22:52Z")
Thanks. Just to be sure, you are only offsetting in the denominator of c\_i=w\_i/m\_i, right?
---
### Author:  [@bull](https://community.drivendata.org/u/bull)
#### Post date: [September 4, 2018, 3:37pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/11 "2018-09-04T15:37:02Z")
Yep. Then the true value will be close, but not equal to zero.
---
### Author:  [@sangxia](https://community.drivendata.org/u/sangxia)
#### Post date: [September 4, 2018, 4:06pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/12 "2018-09-04T16:06:36Z")
Hmm… still a bit confused. Could you write down the evaluation formula with epsilon in it?
I think it matters because as I understand it now, getting one of those small values even a little bit wrong can be quite literally disastrous, even though they only account for a negligible fraction of the values.
---
### Author:  [@ashutoshshukla2894](https://community.drivendata.org/u/ashutoshshukla2894)
#### Post date: [September 4, 2018, 5:26pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/13 "2018-09-04T17:26:03Z")
Okay! Sounds good. One more query, even in the submission format, the temperature data is missing for some entries. That means, temperature need not be predicted, right ?
---
### Author:  [@bull](https://community.drivendata.org/u/bull)
#### Post date: [September 4, 2018, 5:36pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/14 "2018-09-04T17:36:58Z")
Correct, you only need to predict consumption.
---
### Author:  [@ironbar](https://community.drivendata.org/u/ironbar)
#### Post date: [September 9, 2018, 6:57am UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/15 "2018-09-09T06:57:33Z")
Hi,
I have two more questions related to leaderboard:
1. Which percent of the test set belongs to public?
2. The split between public and private test set is random?
Thanks
---
### Author:  [@gijskoot](https://community.drivendata.org/u/gijskoot)
#### Post date: [October 10, 2018, 10:30am UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/16 "2018-10-10T10:30:22Z")
In our submission, for the daily and weekly predictions, are we supposed to predict the average kW over that period or the summed kWh? So, if the average usage is 20 kW per hour, is the correct prediction on a daily entry 20 or is it 20 times 24, 480?
---
### Author:  [@c3josh](https://community.drivendata.org/u/c3josh)
#### Post date: [October 10, 2018, 3:35pm UTC](https://community.drivendata.org/t/about-the-cold-start-energy-forecasting-category/2556/17 "2018-10-10T15:35:40Z")
@gijskoot - answered here [Consumption Aggregation Types](http://community.drivendata.org/t/consumption-aggregation-types/2588/6)