Over the past week or so I’ve been tweaking the OLXPortal source code. Now called just XRD Portal it is quite happily running our little XRD department at the School of Materials, UoM. I’m not quite certain for how long that will be the case since the UoM has bought a commercial project to impose on all facilities (could have had mine for free).
So what has that got to do with stats? Well instrinsic within the design of the portal is its ability to generate usage, user, supervisor and school statistics. BUT what to do with all this fantastic lovely data? This is the main problem one faces with large amounts of data. How to best consolidate and provide in a usable format for digestion?
So what are the new updates?
Firstly I’ve clarried to the users what they need to do in order that they can getting booking permissions on equipment. I had to do this since I was getting emails daily from disgruntled users who felt that by filling in some forms online gave them the right to turn up and start using £100K machines. My favourite is someone complaining “why I had not given me permission”? and then they could not figure out how to open the door or if it was safe to do so! I guess that looks bad on me as there are no bad students only bad teachers using the wrong approach. So I’ve now written everything down so they understand WHY you can not book just yet.
Next I’ve been through and released a code idea I had implemented at the very begining of the project. That of safety deletions. What? As part of the process of using equipment each user must undertake a minimum of two online safety tests. The first is a general test which covers their access to facility things like fire exits and emergency procedures. It also includes simple things like out of hours usage and notification on publications and manuscripts, etc. The next test (which could be more than one) is undertaken with machine specific information. Each machine including a risk assessment as when as SOP and other information. These online tests are now self deleting after a finite period from first undertaking that training. This does not remove the users permissions to book or forget what they have done only suspends them until they complete the training online again.
This also links to upgrades to the online tests which now also include random “stupid” questions and statements which means if a user ticks them they will instantly fail as it is proof they have not read the text. It also includes other safe guards to try and ensure that we have evidence that the user has had to take time review the statements not just tick and agree.
But with these upgrades I could name and shame the students (this is fair definition) that have to undertake the test due to failing because they have not read the text properly. I have lists with some students failing 5 or 6 times for the exact same reason! Should I employ these stats? Should I take those users to one side before triaining them and ask why they kept ticking boxes sayings things like “I promise to bring biscuits to the XRD lab everytime I visit” or “I’m a little teapot short and stout”.
Along with these safety updates I’ve unlocked ics generation and some additional shortcuts used for navigation for the admins and of course implemented some new stats generations code.
Stats - that is short for Statistics (lazy that)
To make sure my boss can promote, defend and justify the XRD department’s existence I’ve been generating stats. Initially these stats were just generated when requested. However, last year I created a system which would provide the supervisor and MSc students’ a usage summary and index of all their data at the end of their project. This system is also rolled out for all users at the end of the academic year and the same time as global “general” permissions wiping. So coupled to these reports I am now generating automonously allocation period reports (AP - yes just like the synchrotron days but now only three months long) as well as six month reports as well.
These reports include:
- New Registered Users
- Machine Usage
- by school
- by role
- by supervisor
- by experiment type
- machine operating times
- the above broken down by machine
Well that is enough for now. The only question(s) now is “Should there be more stats” and “Should these stats be public”?