One of my favourite parts of attending the ever-growing Alteryx Inspire 2016 conference and its like is hearing about the fun new features that tools such as the wonderful Alteryx are going to make available soon. It’s always exciting to think about how such developments might improve our job efficiency, satisfaction or enable whole new activities that so far have not been practical.
From this blog’s page view stats, it seems like others out in the great mysterious internet also find that sort of topic interesting, so below are a few notes I made from the various public sessions I was lucky enough to attend, about some of what Alteryx is thinking to add over the next few versions.
Since the addition of in-database tools, Alteryx has allowed analysts to push some of the heavy lifting / bandwidth hoggage back to the database servers that provide the data to analysts. If you’re an analyst who regularly uses moderate to large datasets obtained from databases you should really look into this feature, as by default Alteryx spends time sucking data from the remote database to your local machine otherwise. Anyway, a few new developments are apparently planned:
- New in-database data sources.
- New in-database predictive analytics (I believe SQL Server was on the list)
- A makeover of the in-database connection tool to make it easier to use
New data sources:
- The Adobe Analytics connector was recently released (here’s how to use it) – they plan further refinements to it. Nothing in particular was specified, but users are invited to send in their ideas.
- A connector for Salesforce Wave.
New predictive tools:
Many of these may be delivered via the Alteryx Predictive District (in fact it’s very worth looking there now for the existing tools – although I appreciate they don’t want to clog up your toolbar with thousands of icons, it’s not always easy to remember to check these fantastic districts! May I suggest an in-Alteryx search feature for these in the future?)
- Time series model factory
- Time series forecast factory
- Time series factor sample
- Cross validation model comparison
- Model based value imputation
- K medoids cluster analysis
- Text classification tools, to enable e.g. sentiment analysis, key phrase extraction, language detection, topic modelling.
An analytic app that will allow you to install your own choice of R packages from CRAN.
Some “Getting started kits” that will help newcomers to predictive analytics, each focusing on a specific business question, examples include:
- How does a price change impact my bottom line?
- How can I predict how much a customer will spend?
- How can I predict whether a customer will buy the produce I put on sale?
Prescriptive analysis tools:
Yes, the next stage after predicting something is to prescribe what we should then do. A new toolbar will come with tools in this category. Starting with:
- Optimisation: have Alteryx maximise or minimise a value based on constraints for an optimum outcome. One example demonstrated was “what’s the best product mix to stock on a shop shelf to maximise profits, whilst ensuring the shelf has no more than 1 of any particular item?”.
- Simulation: think here of things like Monte Carlo simulations, and, in the future, agent-based simulations.
Improvements to existing tools:
- Formula tool: will include
- inline search for functions & fields,
- suggestions of common options based on context such as field type,
- a data preview to show you right away what the results of your formula will be on a sample record.
This one makes me happy! Without meaning to cause offence, the current incarnation of the formula tool, which has to be one of the most used tools for most everyone, is a little…erm…”old fashioned” to those of us spoilt with auto-correcty/lookup things from other vendors in recent times when typing in code. No more digging around trying to remember if a function to create a date is under “date/time” or “conversion” etc.
- Smarter data profiling tools
- Improved reporting output tools
- Web based scheduling
Alteryx server updates:
I must admit to not being a server user, so I am not 100% these whether these are new features. But it seemed so:
- Row level security on data, i.e. different users see different records in the same datasource.
- Version history
Estimated release dates:
Version 10.6 may be around the end of this month. Version 11 towards the end of the year (no promises made). I did not note which features were planned for which version.
3 thoughts on “Future features coming to Alteryx 10.6 and beyond”
Wow, wow, wow! Pure awesomeness. Can’t wait!
Do you know if Alteryx has an imputation tool for seeding missing values for a Time Series problem?
I have looked at the Imputation Tool which is a general-purpose tool and not suitable for imputation required for Time Series problem.
Hi – I’m not aware of a tool specifically for that I’m afraid.
I do know there’s a “TS Filler” tool that will fill in the missing dates from a time series – https://help.alteryx.com/2018.2/TimeSeriesFiller.htm
But that’s more to fill gaps and doesn’t impute.
I saw a post on Alteryx knowledge base that might be relevant, depending on how you wish to impute those values? Impute missing data values with a custom formula: https://community.alteryx.com/t5/Alteryx-Knowledge-Base/Impute-Missing-Data-Values-with-a-Custom-Formula/ta-p/87312
I’m sorry I don’t know a simple tool answer 🙂 but good luck!