Sybil.dev

Sybil.dev

Forecast anything with one simple Google Sheet formula

4 followers

Visit website
With Sybil, you can predict any time series, leveraging our powerful forecasting algorithm in one simple Google Sheet formula. ā € Sybil can predict non-linear trends, taking into account seasonality, shifts in the trend, and even outliers!
Sybil.dev gallery image
Sybil.dev gallery image
Sybil.dev gallery image
Launch tags:Web Appā€¢Productivityā€¢API
Launch Team
Anima - Vibe Coding for Product Teams
Build websites and apps with AI that understands design.
Promoted

What do you think? ā€¦

Chris Davis
I think meeting people where there are (in spreadsheets) by creating a custom formula is ingenius! Kudos
Pierre
HappySteps: Mental Health Help
Maker
@imchrisdavis thank you, that's exactly our angle :)
Pierre
HappySteps: Mental Health Help
Maker
Hello team PH, maker here together with @carbonaro šŸ™ƒ šŸ™ First, many thanks to @chrismessina for hunting us! šŸ’” The idea behind Sybil.dev is fairly simple: what if you could leverage sophisticated forecast algorithms with a simple google spreadsheet formula? šŸ“Š That's exactly what this is all about. In under 2 minutes you can set up a forecast for : - stocks - production volumes - sales - prices - temperatures - whatever ^^ šŸ“ˆ This works basically for any time series, and takes into account micro / macro trends, embedded seasonality and much more. Together with @carbonaro we're happy to answer any question you may have!
Joe
@carbonaro @chrismessina @pierrestanislas Do you have common examples Screenshots
Nicolas Le Roux
PrivacyAngel
How can I use this with monthly time series, not daily ones? Thanks.
Pierre
HappySteps: Mental Health Help
Maker
@nico_lrx you can feed the model monthly series, and then extract the relevant monthly points from the daily output (disregarding whatever's in between). See our sample spreadsheet which does exactly that on the aggregated chart tab using vlookup : https://docs.google.com/spreadsh...
Marc Dupuis
@pierrestanislas, this is cool and could be really powerful. Without revealing the secret sauce, could you roughly describe how this is powered? Is it some sort of LSTM model behind the scenes?
Marc Dupuis
@pierrestanislas Follow up question: are you thinking about a way to identify correlation between two separate trend lines? I work in the enterprise forecasting space and can elaborate a bit more on the use case and value if you're interested.
Pierre
HappySteps: Mental Health Help
Maker
@marc_dupuis2 Hello - regarding your first question: yes, there is some LTSM theory behind this, plus custom made transforms to extract specific seasonalities at various scales (sorry, can't say much more ^^). Regarding your second question : we'd be happy for you to elaborate a little bit, not sure we understand?
Marc Dupuis
@pierrestanislas using timeseries, you may want to track correlation between two events in time. So for example you may want to know if thereā€™s a correlation between marketing spend and customer acquisition. So the customer acquisition forecast is not only a factor of historical trends of that metric, but also a factor of additional variables that may be related.
Klim Yadrintsev
@pierrestanislas @carbonaro Is it only day interpolation? What if I want to use monthly data as a source and also make predictions based on that?
Pierre
HappySteps: Mental Health Help
Maker
@carbonaro @klim_yadrintsev Hi, you can feed the model monthly series, and then extract the relevant monthly points from the daily output (disregarding whatever's in between). See our sample spreadsheet on our website, which does exactly that on the aggregated chart tab using vlookup.