Is anyone out there integrating Tessitura Analytics with other software? Anyone working on a project that will do this? I haven't made any of the virtual coffees, so I'm afraid I don't know the current state of play..
Full disclosure: I'm working with the planning committee to develop a session specifically about this, and we are still looking for potential presenters.
Gawain Lavers you can definitely integrate the system underlying Tessitura Analytics (Sisense) with R. In addition, there are a number of third party sisense and community plugins that can be used with Sisense. You can also integrate Sisense with Amazon Echo. I've investigated a number of these things.
However, once I get our environment up and running I discovered that because I'm on a Shared Instance of Sisense on RAMP. I can not make any of these changes. Only Tessitura Consulting staff are allowed to do customizations on RAMP Sisense. I investigated getting off of a shared Sisense. And discovered that the price to do this was going to be prohibitively expensive. So all of those options are not on the table for me. We have no budgeted extra for private systems or Consultant customizations.
I'm going to spend my effort getting data out of our database to do local data science activities. This was one of the big disappointments for me over the last year.
Regarding integrations of Tessitura Analytics with other software. You are most likely to have to look at the big organizations. I'd check with folks like Samuel Tran over at AMNH. Or other folks who have onsite instances.
--Tom
Hi Gawain Laversand , I ingested some weather data into the very first alpha version of Tessitura Analytics in 2016 in order to correlate attendance with weather data (Machine Learning and python). We were able to develop a predictive model for attendance forecasting using weather forecast data. Unfortunately I haven't had the chance to work with our latest Tessitura Analytics sandbox. However, we recently developed an exhibition recommendation system using ticket purchase history (stored in a data warehouse) and machine learning (python). Still in the experimentation phase though.
Hi Gawain,
I know of a handful of Analytics customization projects already happening. I’m not sure if they fit the profile you’re looking for, so I’ll email you directly to chat about it.
We’ve advised some members to hold-off on customizations (if feasible) because we’re testing a new Tessitura Analytics customization framework that enables custom Analytics dimensions and cubes via system tables. This feature will make it much easier to integrate custom data from the Tessitura DB to the cubes in a way that will be sustainable and maintainable, even surviving Tessitura upgrades.
This is especially exciting in RAMP where we prohibit direct access to the cubes in order to maintain our excellent security standards. We’re starting by exposing this new functionality only to Tessitura Consulting to ensure we work out all of the kinks. Our long-term goal is to release it to everyone.
John Jakovich said:We’ve advised some members to hold-off on customizations (if feasible) because we’re testing a new Tessitura Analytics customization framework that enables custom Analytics dimensions and cubes via system tables.
Hmmm. That's less than ideal news for our Conference Session...
I had thought, however, given Analytics' newness, and the fact that it's a quickfire session might give us the opportunity to present on "planning stage", or in Samuel's example "did work". Thanks for reaching out!
Hey Tom,
Our team is taking a similar approach. We see Analytics as a replacement to T-stats and a more user friendly ad-hoc reporting tool. But for our BI team, we have been spending some time pulling data out of the database and doing R and Python notebooks for analysis.
@grantoffermann2464
I'm beginning to think about Tessitura Analytics in exactly the same way you are. Tessitura Analytics is a T-Stats replacement for advanced users. (Which by itself is Very Cool). But, in its current form is not an approach to front end advanced analytics, like churn modeling, GIS, forecasting..
You might be interested, I've got working code that can pull Tessitura Lists and output sets data through the REST API into an R based Jupyter Notebook. If you are interested let me know. Talked a little bit about this at last Friday's Build this about two weeks ago. The idea is that you create a list and output set in Tessitura V15. Then use R's httr library to pull JSON form the REST API to pull the content into a local store to do your analytics. This works fairly well for me. Laura Burgos (Past Member) and I have been discussing this as an interesting approach for getting data for advanced analytics.
We are also working on a Python Library over in the Developers Slack Channel. This is not yet able to pull long Tessitura lists/output sets. But, I'm working with the team over there toward that outcome. Once we have that you could do a pip install tessitura and then make calls to the REST API for list and output set content as the basis of your analysis.
Please let me know if you or your team are interested in discussing this further.