If you’re a researcher working on electricity disaggregation (aka NALM / NILM / NIALM) then would you find a wiki and/or online community useful?

Background

I’m currently doing a Ph.D. on “smart meter disaggregation”. I’m in a lovely and very supportive department and research group but no one else in my university is doing disaggregation so I find myself having to spend time gathering information which I’m sure all disaggregation researchers have to gather. This feels rather inefficient. And I’m not alone: my understanding is that many disaggregation researchers find themselves in situations where they are the only ones working on disaggregation in their institution. Instead of working in these isolated little “islands of disaggregation”, wouldn’t it be better if we had a wiki and/or online community where we could pool our knowledge? This might allow new researchers to get to the “meaty” work as quickly as possible; and would allow seasoned researchers to efficiently keep abreast of developments.

Wiki

Some things we could use the wiki for:

  • Detailed descriptions of appliance power consumption. Each appliance “article” would have several plots of the appliance waveform (from different datasets), describe each power state, histograms of appliance usage over an average day etc. Highlight differences between different datasets (e.g. US washing machines look nothing like UK washing machines).
  • List of conferences, workshops, summer schools and journals suitable for NALM researchers
  • Table of research. Each row would have fields such as “sample rate”, “measured parameters (active power / reactive power / voltage)”, “classifier (HMM / NN / SVM / KNN / DT etc)”. The main purpose would be to make it easier to quickly find, for example, “all recent NALM research on *1Hz active power data”.
  • Table of NALM competitions and winners.
  • List of data collection hardware
  • Software resources (NALM code, lists of HMM toolkits etc)
  • List of energy recommendation projects
  • Proposed metadata ontology for energy data
  • List of NALM blogs
  • Table of public datasets. For each dataset, have fields for “DOI of paper(s) describing dataset”, “duration”, “sample rate”, “measured parameters”, “number of houses” etc etc

Software for running the wiki

In terms of software for running the wiki, I’m currently thinking along the lines of MediaWiki (the software which runs WikiPedia) with the Semantic Media Wiki extension (and others).

Online community

As well as a wiki, it might also be useful to have an “online community”. This could be as simple as an email list or a Google Plus community,

  • interesting new papers (including our own ;) ). An online “reading group”.
  • technical questions (e.g. “I’ve noticed something odd in dataset X; has anyone else had this problem?”)
  • Decide on a metadata ontology for energy data? (UKERC EDC have one but it lacks things like sample rate (different for appliances and mains) etc)
  • debating techniques (“I think my favourite machine learning approach is a great choice for disaggregation. Why hasn’t anyone used it for disaggregation yet?”)
  • Scoping out BSc / MSc / PhD (e.g. “I want to propose an MSc project where the aim is to write an open-source implementation of the entirety of Hart’s work. Has anyone done this already?”)
  • companies / labs looking for collaborators / internships / employees in disaggregation
  • shamelessly linking to our own blog entries on disaggregation ;)
  • announcements of disaggregation competitions (I wrote some notes about NALM competitions on Oli’s Blog)
  • announcing / organising NALM workshops / conferences / meetups etc

It takes a while to start communities

Of course, it necessarily takes a while for any wiki / online community to “warm up” and reach critical mass. I’m entirely expecting to be the only one posting content to the wiki for a while! That’s fine (although it’d be awesome if it wasn’t the case!)

More details

My detailed (but scruffy) notes about the wiki and online community are in this Google doc. And here’s the Twitter conversation about this blog post.

Specific questions!

Some questions that it would be really good to discuss in the comments…

(Please see this google doc for existing responses to some of these questions)

  • would you find a wiki useful?
  • what about an online community?
  • domain name for the wiki and/or community?
  • software for running the wiki? (Semantic Media Wiki? Drupal? Something else?)
  • platform for the online community? (Google Groups? Google Plus Community? Self-hosted mailing list e.g. GNU Mailman? Self-hosted web forum on the same domain as the wiki?)
  • ideas for content / content structure?