Smart Predict #ASUG Webcast Recap with Q&A
Intro – 0:00
Demo – 12:46
Roadmap – 37:37
Q&A – 47:36
Don’t forget to join us for more SAP Analytics Cloud upcoming webcasts:
December 4 INFL: Mobile BI Influence Council Re-launch
December 12 BI: SAP Analytics Cloud – Connectivity Deep Dive
January 30 BI: SAP BW/4HANA 2.0
January 31 BI: SAP Analytics Cloud Roadmap
The legal disclaimer applies; things in the future are subject to change.
We are want to adopt machine learning, but struggle to do this, and we do not know where to start. The stigma is with advanced analytics being too complex, with only data scientists being able to achieve meaningful outcomes. Why is this? Overly complex and geared towards “Data Science” personas; required learning specific algorithms or coding languages and challenging user interfaces and workflows
The opportunity is there as Gartner expects 50% or more of advanced analytics to be done by ‘citizen data scientists’, Business Analysts significantly outnumber ‘Data Scientists’ and are much more cost effective, and capturing even a small % of the business analyst community drives larger user growth than the Data Science community
With “intelligent enterprise” era, organizations need to have seamless integration between all applications and technology to be able to act quickly. SAP Analytics Cloud brings together BI and Planning and enhances those existing capabilities with Machine Learning and AI, with one platform providing augmented analytics to help you act confidently
SAP Analytics Cloud guides business analysts on their journey to Smart Analytics by exposing analysts to machine learning concepts gradually through existing BI and planning workflow, users of any skill level. The first set of augmented capabilities “Search to Insight” and “Smart Insights” helps users alike obtain answers faster out of their data, with how we would normally search for new information: by asking questions.
Instead of having to think about which visualization would be best to help you understand your data, you can use natural language to query your data, and have visualizations and additional explanations created for you, on the fly, delivering answers to questions without the need to build visualizations manually.
Another set of augmented capabilities includes Smart Insights and Smart Discovery to help explain data point, measure, or dimension. If you want to understand more about what drives the value of sales opportunities, or revenue, or expense, maybe compare one group of customers against another, SAP Analytics Cloud will conduct the analysis for the analyst to provide key influencers and drivers of a particular outcome and the solution will also build out an entire story, highlighting important relevant data points through visualization. For analysts, they do not have to start from scratch, and they can continue to enhance that story with custom insights.
Smart Predict offers the most advanced set of augmented capabilities and is focused on showing the next best action to take through artificial intelligence. After business trends and patterns have been understood, Smart Predict allows business analysts to create a plan on where to focus, and who to target.
Citizen Data Scientists (business analysts) are predicted to create more advanced analytics insights than any other user. These business analysts are the future, and SAP Analytics Cloud is designed to guide these users from traditional data discovery through to Smart Analytics.
See the replay for a demo
Smart Predict was released end of Q2
Gradually on-boarding customers and working on use cases
SAP Analytics Cloud is moving to a quarterly release cycle; more to come.
Instead of incremental improvements on a bi-weekly basis, moving to quarterly
Some things in Q4 have moved to Q1 (planned for February)
Smart Predict works on spreadsheets uploaded; and acquire OData through S/4HANA; in Q1 want to extend acquisition support to other SQL datasets and SAP HANA. Enable acquisition in a more governed way.
Extend to SAP Cloud/on-premise applications through an OData connector.
Continue to simplify user experience
Different ways to simplify; over time like collapsing 2 step process in a single operation
Plan to publish scenarios in S/4HANA.
Allow exporting predictive datasets to integrate with custom apps and database
Focus on Analytics Cloud Network – invoice forecasting in Q1 (planned)
Want to democratize predictive analytics
Planned for Q2/2019 – most important, support HANA live data connections, start with SQL views and tables, with more artifacts over time. Plan for automated data preparation (on premise had Data Manager, looking for simpler way in cloud).
Plan to have scheduling capabilities to refresh predictions, refresh models over time.
|which smart features are supported withlive connection?||Please visit this wiki for a comparison: https://wiki.scn.sap.com/wiki/display/BOC/SAP Analytics Cloud Support Matrix|
|when will predictive scenario be available to customers which already have planning and predictive license?||It is available for those with a BI license; if you have an existing tenant it will be updated in 2019|
|Is it planned to get this workung with more then offline data models? especially HANA Live Connection would be great…||He has a roadmap slide at the end…let us know if that doesn’t answer it.|
|how is the best number of column /measure to do a significant analysis?||the more rows you have, the most robust the model. Not less than .95|
|Does the Smart discovery work when you have hierarchies setup within your data model?||don’t know…will follow up|
|The smart view created a list of ‘outliers’ based on expected values. How are these expected values calculated?||when you do Smart Discovery, build model to explain variable; within model most records correspond…records that don’t correspond are outliers|
|which license do I need to use Smart Predict?||BI license for SAP Analytics Cloud|
|what the technical difference between the Dataset and Model objects?||Dataset is data and model is your SAP Analytics Cloud model.|
|will Value driver tree also be possbile with HANA Live data connection in %2BQ2/2019 as mentioned in the slides?||will check|
|can I connect a dataset to a source system as in live connection?||see roadmap; planned for q2/2019|
|Is this ROI simulation standard? or was it modeled specifically to this purpose?||simulation is standard|
|Are these smart functionalities a separate license or part of the BI license?||part of BI|
|Hi, when do the partners tenant get the smart predict capability. ?||You can provision a new tenant today or wait to update in 2019.|
|Can we get the sample data that you have used during the demo?||Please see these blogs on SAP Community for examples – https://people.sap.com/sarah.detzler|
|For the predictive scenario, what is the source that you used?||used export from CRM of opportunity data|
|Great presentation. I liked the demo. I am an existing SAC customer. Does my tenant has the smart predict feature available? If not, what I need to do to get it. Is there an additional license beyond Planning and Analytics feature for us to use this feature?||You can ask for a new tenant to be provisioned or wait until 2019; contact your SAP Customer Success rep|
|What’s the timeline to integrate this with BW on HANA for live connectivity?||Q2/2019 – for HANA live – see his slide now|
|Oh ok.. I though that’s related to native HANA connection and not BW on HANA. Thanks for clarifying.||I said they are working on it, no time commitment|
|will new predictive features come in SAC only quarterly in 2019?||QRC is a different – effective SAC will be quarterly cycle|
|Can you clarify on licensing question?||what’s the question? This comes with the BI license|
|future quarterly release schedule – is this for all SAC – or just SAC smart predict?||SAC|
|What are the data limit to run Smart Predictive||currently 1K columns 1m rows|
|What % of Predictive Power would be for the first predictive scenario? Will it start from 0%?||Predictive accuracy gives % of target correctly – 0% – nothing, 70% – model 70% of targets correctly – power – how confident work 95%|
|When include more models like Clustering , Association, etc||beyond Q2 next year|
|Based on the Roadmap is the intention to replace on prem-PA?||on premise PA will be supported and sold; will on premise is interested as it supports Data Scientist and Business Analyst workflows for large scale operations|
|Is Live connection for VORA foreseen?||2 ways -interop with Data Hub 2 – support at some point HANA Live SDI SDA|
|Will it get possible to write back simulations to a normal HANA DB….maybe in a specific table?||plan for that apply for workflow live|