User Group Meeting at PI World 2018 - Opportunity to share thoughts on topics of interest to Transportation

As a follow up to our meeting at PI World 2018 in San Francisco, below are the meeting notes and ideas shared by the participants (17 people were in the room).

Thanks to all who attended and contributed! Let’s keep the discussion going and start kicking off some actions from the list below:

Analytics Applications Map by Expected Outcomes

Users would like to leverage the community to map Dataset + Expected Outcome to the appropriate Analytic Technology that would be best suited.

Data Sharing for Analytics, Benchmarking, and Monetization

- Partners are looking for anonymized datasets by asset class to be shared by users with the community so that they can be used as starting point to explore potential underlying value of applying advanced analytics.

- The community could also collaborate on sharing data sets by asset classes which would allow benchmarking asset performance by manufacturer / operating conditions. 

- The user group could elect sub-committees to define and control how data sharing among the group could take place. This may be a combination of process and technologies (e.g. Block chain, etc.) with the end goal of identifying ways for users to monetize their shared data.

- The idea of having OEM sell assets and also ship optimal/benchmark operational data profiles which could be directly compared to actual operations would also bring value to users in the group.

-It was suggested to initiate data sharing collaboration on non-competing sides of the business such as safety. Users may be more willing to share data if it is meant to increase safety of operations rather than other aspects of their business.

Data Quality

- Users and partners are interested in having established data quality metrics, possibly built-in into the data infrastructure. For instance, if a sensor could have an associated quality score/metric/KPI, that would make it much easier to know what data can be used in analytics, reporting, and where problems need to be fixed.

- One idea to get started would be to collaboratively define ‘Data Quality’. Participant suggested knowing some scenarios having been shared in the past, some being quite simple to identify, but some are more difficult (e.g. sensor drift).

- The concept of data quality ‘Classification’ was suggested: or the application of “feature extraction” principles to identify quality issues – allowing users to decide how to handle identified questionable data.

Data Access by Asset Class

Users are interested in crowdsourcing a list of tools and techniques to connect to various pieces of equipment / assets. The idea comes from the fact it is sometimes a challenge to find a way to 'connect' to some data sources, and leveraging prior experience from others would allow all to save costly trial and errors and research.

Scalability and Performance Scenarios

The user group could collaborate on identifying scalability and performance scenarios which would need to be supported now and over time so that they can be tested and reference architectures can be proposed.

Software Improvement and Enhancement Requests

Users have voiced some recurring software enhancements requests and suggested a way for the users group to unite to ass weight on some lingering issues they would like to see fixed.

As a follow up to our meeting at PI World 2018 in San Francisco, below are the meeting notes and ideas shared by the participants (17 people were in the room).

Thanks to all who attended and contributed! Let’s keep the discussion going and start kicking off some actions from the list below:

Analytics Applications Map by Expected Outcomes

Users would like to leverage the community to map Dataset + Expected Outcome to the appropriate Analytic Technology that would be best suited.

Data Sharing for Analytics, Benchmarking, and Monetization

- Partners are looking for anonymized datasets by asset class to be shared by users with the community so that they can be used as starting point to explore potential underlying value of applying advanced analytics.

- The community could also collaborate on sharing data sets by asset classes which would allow benchmarking asset performance by manufacturer / operating conditions. 

- The user group could elect sub-committees to define and control how data sharing among the group could take place. This may be a combination of process and technologies (e.g. Block chain, etc.) with the end goal of identifying ways for users to monetize their shared data.

- The idea of having OEM sell assets and also ship optimal/benchmark operational data profiles which could be directly compared to actual operations would also bring value to users in the group.

-It was suggested to initiate data sharing collaboration on non-competing sides of the business such as safety. Users may be more willing to share data if it is meant to increase safety of operations rather than other aspects of their business.

Data Quality

- Users and partners are interested in having established data quality metrics, possibly built-in into the data infrastructure. For instance, if a sensor could have an associated quality score/metric/KPI, that would make it much easier to know what data can be used in analytics, reporting, and where problems need to be fixed.

- One idea to get started would be to collaboratively define ‘Data Quality’. Participant suggested knowing some scenarios having been shared in the past, some being quite simple to identify, but some are more difficult (e.g. sensor drift).

- The concept of data quality ‘Classification’ was suggested: or the application of “feature extraction” principles to identify quality issues – allowing users to decide how to handle identified questionable data.

Data Access by Asset Class

Users are interested in crowdsourcing a list of tools and techniques to connect to various pieces of equipment / assets. The idea comes from the fact it is sometimes a challenge to find a way to 'connect' to some data sources, and leveraging prior experience from others would allow all to save costly trial and errors and research.

Scalability and Performance Scenarios

The user group could collaborate on identifying scalability and performance scenarios which would need to be supported now and over time so that they can be tested and reference architectures can be proposed.

Software Improvement and Enhancement Requests

Users have voiced some recurring software enhancements requests and suggested a way for the users group to unite to ass weight on some lingering issues they would like to see fixed.