Aberdeen Air Quality

Update: A write-up of this event which took place on 16-17th February 2019 is available on this page.

How much do you care about the quality of the air you breathe as you walk to work or university, take the kids to school, cycle or jog, or open your bedroom window?

How good is the air you are breathing? How do you know? What are the levels of particulates (PM2.5 or PM10) and why is this important?

pm25_comparison
pm25_comparison

When do these levels go up or down? What does that mean?

Who warns you? Where do they get their data, and how good is it?

Where do you get information, or alerts that you can trust?

We aim to sort this in Aberdeen

Partnering with community groups, Aberdeen University and 57 North Hacklab, we are working on a longterm project to build and deploy community-built, and hosted, sensors for PM2.5 and PM10. We aim to have fifty of these in place in the next few months, across Aberdeen. You can see some early ones in place and generating data here.

The first significant milestone of this will be the community workshop we are holding on 16-17 February 2019. If you want to be part of it, you can get a ticket here. But, be quick; they are going quickly.

Weekend activities

There are loads of things you can do if you attend.

Sensor Building

For a small cost, you can come along and build your own sensor with someone to help you, and take it home to plug into your home wifi. It will then contribute data for your part of the city.

But we will be doing much more than that.

Working with the data

If you have experience in data science or data analysis, or if you want to work with those who do, there are loads of options to work with the data from existing and future sensors.

These include

  • Allow historical reading to be analysed against the official government sensors for comparison
  • Use the data; wind speed, humidity… to build live maps of readings to identify sources of emissions.
  • Compensate readings from sensors against factors which affect pollution levels to attempt to understand the emissions of pollutants in a given area.
  • Build predictive models of future pollution
  • Fix a minor issue with the existing data Collected Data (see https://github.com/opendata-stuttgart/madavi-api/issues/8 )
  • Build an API for the access of the Luftdaten sensor data to allow querying of the sensor data

Software development

If you are a software developer or studying to be one, you could

  • Create alerts systems to warn of anticipated spikes in pollutants, perhaps using Twitter, or email.
  • Add to the code for the Luftdaten sensors to allow connection over LoRaWAN interface.
  • Create LoRaWAN server code to allow sensors to feed up to the Luftdaten website.
  • Security testing of the IoT Code used by the Luftdaten sensors.

Community Groups / Educators / Activists / Journalists

You don’t have to be a techie! If you are a concerned citizen, and community activist, a teacher, or a journalist there is so much you could do. For example:

  • How can you understand the data?
  • Identify how this could assist with local issues, campaigns, educational activities.
  • Help us capture the weekend by blogging, or creating digital content

Even if you just want to be part of the buzz and keep the coffees and teas flowing, that is a great contribution.

See you there!

Ian, Bruce, Andrew and Steve

Header image by Jaroslav Devia on Unsplash

2018 – A year in review

2018 has been a really busy year for us. Here are all the things that we delivered.

Open Data Camp

We hosted UK Open Data Camp’s first ever visit North of the border in November. Over a hundred people travelled to Aberdeen for two days of unconferencing where there were 44 sessions run on a variety of data-related topics. Some people went for an Aberdeen version of the Joy Diversion walk around old Aberdeen, and others discovered the pleasure of logging Open Benches. The feedback was overwhelmingly positive and there were loads of write-ups.

Code the City Hack Weekends

We had two great Code The City Events: CTC13 – Hacking our Relationship with Alcohol, and CTC14 – Archaeology. Both were well attended and produced some very interesting results. The first saw us tackling some interesting real-world problems, helping people to overcome problems, and build a machine learning model to predict whether a beer can design would be more likely to be perceived as alcoholic or not. A report of the weekend is being written as an academic paper for a forthcoming health conference!

 

The second weekend saw us scanning and creating 3D renders of six real skeletons with mobile phones. We also began to create a 3D model of the church in which the dig took place, and generate data from written logs to populate that.

Well done to all who participated. We got some great feedback on each event.

Data Meetups

Wearing our ODI hats, we launched the new monthly Data Meetups in April – and managed to squeeze in nine of them this year. These are really well attended, and saw over 300 people in total coming out on a Tuesday night to hear speakers from across the country on a diverse range of data topics. These ranged from Creating a Data Culture in your business, to public Open Data; from the data of Scottish Football to the use of blockchain in Oil and Gas; and from the use of IoT in Agriculture to extracting data from photos published on Flickr in order to assist conservation.

Open Data

We’ve also been lobbying the Scottish Government and the city council on Open Data, as Ian has been writing on our sister site. That is starting to bear fruit. Aberdeen City Council have soft-launched a new open data platform, and are recruiting a manager for their open data work. While this is good, it is not as impressive as Dundee and Perth‘s new platforms, yet. The Scottish Cities Alliance are recruiting a new programme manager, and Ian has been invited to be part of a round table discussion on the way forward for Open Data hosted by the Scottish Government next February. It sounds like things will start to move in the right direction in 2019!

Research

Ian and Andrew have worked with ODI HQ to run two local workshops, contributing to two national pieces of research: the first on the effects of Peer to Peer markets on accommodation, and a second on what barriers there are to the better use of Ordnance Survey data and services.

Here’s to an equally successful 2019! Have a great festive break folks!

Ian, Andrew, Steve, Bruce

 

When data glitches ruin your day

The Scotrail service between Aberdeen and Inverness is currently disrupted by some improvement works to the line between Aberdeen and Dyce. Bus replacement services are in operation between Aberdeen and Dyce, with normal service between Dyce and Inverness. All because the line has been dug up in preparation for new dual tracks being laid.

But if you do a search on the Scotrail website for the next train north from Aberdeen you get this message:

Train_Tickets___Times___Timetables___Fares_in_Scotland___ScotRail

“No services between these stations have been found”. One would assume that Scotrail isn’t running services from Aberdeen to Insch. The same thing happens for searches to all stations north of Aberdeen. Continue reading When data glitches ruin your day

Cantcha @ codethecity 13

During Codethecity 13, a project arose around the clarity of labelling of alcohol, initiated following a presentation from @wayne_gault about alcohol labelling and marketing.

Quickly the focus of discussion moved to craft beer, and the rising importance of graphic design in the branding and positioning of the various craft breweries, often using cartoon and other colourful illustrative styles. In a competitive market with many new entrants it’s no surprise that some corners are being cut. Sometimes through a lack of awareness of the issue, sometimes not so much.

Sending one particularly colourful can to the Google Vision API was telling. Google identified it as a “Yellow, Aluminium Can, of Soft Drink”. You can try the vision API yourself here.

Vision_API_Google_Cloud

The team working on this issue came up with the idea of creating a new version of reCaptcha. Rather than clicking on all the store fronts, or vehicles, you would click on all the beers, or all the energy drinks.

This would both raise awareness of the issue, and provide a data set to demonstrate the level of confusion around specific labels. If the ‘spongebob’ beer in the mockup below is regularly identified as an energy drink, perhaps the labelling falls short of the standard.

Screen Shot 2018-06-09 at 19.47.22

How this could be useful

The Portman Group publishes the Code of Practice for the naming, packaging and promotion of alcoholic drinks. An ultimate goal of the CANtcha team was to use data generated from the app to add weight to submissions to the Portman Group where packaging falls short of the Code.

Some early (and far from rigorous given the timescales) testing showed that enough of the alcohol was being misidentified that an issue likely exists.

Code for the mock ups, some initial data sets, and some initial analysis are available on Github. A more technical write up is also available.

Even though this was a very brief, short lived prototype – I’ll never look at a beer in the same way again

I subscribe to a craft beer delivery service. It sends me a case every now and then with enough interesting beer from around the world to keep my ‘bottle of hipster a week’ habit going.

A case just arrived.

Now that I’m aware of the Portman rules, I couldn’t help wonder what CANtcha would make of a couple of them:

IMG_20180607_193824

IMG_20180607_193738

I suspect both of these might fall short on one or more counts. I also love both of them. The Google API mentioned above identifies one as beer, and the other as an energy drink. It’s likely obvious which is which.

Neither is identified as a can of coffee, surprisingly.