Numercon yesterday was awesome! Hope and I were able to get across what we are building with NumerBlox. The presentation showed a list of next steps we are excited about. We ticked off 2 of them for NumerBlox v0.2.8.
NumerBlox presentation:
Conversation
Replying to
1. FNCv3 and Exposure Dissimilarity metrics in Evaluators. FNCv3 (mean, std and sharpe) is now calculated if you use NumeraiClassicEvaluator. Exposure Dissimilarity can be calculated for both Classic and Signals.
1
2
2. numerbay.ai integrations! Big thanks to ResTrading for implementing a way to pull bought predictions (NumerBayCSVs) and to let sellers upload predictions (NumerBaySubmitter).
NumerBayCSVs:
crowdcent.github.io/numerblox/mode
NumerBaySubmitter:
1
4
3. Fix bugs in NumeraiClassicDownloader in anticipation of Numerai v4 dataset. The Downloader should be easily adaptable to the now more frequent dataset updates we can expect from Numerai. We feel prepared to adapt it when the v4 data release is launched.
1
3
NumerBlox (pip install numerblox)
pypi.org/project/numerb
Github:
github.com/crowdcent/nume
Documentation:
crowdcent.github.io/numerblox
RocketChat:
rocketchat.numer.ai/channel/numerb
2