Code to support the analysis of the AAMOS-00 anonymised dataset.
You can find further details about this dataset in the publication "Home Monitoring with Connected Mobile Devices for Asthma Attack Prediction with Machine Learning" - Scientific Data, Nature.
We carried out the 2-phase, 7-month AAMOS-00 observational study to collect data about asthma status using three smart-monitoring devices (smart peak flow meter, smart inhaler, smartwatch), and daily symptom questionnaires. Combined with localised weather, pollen, and air-quality reports, we collected a rich longitudinal dataset to explore the feasibility of passive monitoring and asthma attack prediction.
Between June-2021 and June-2022, in the midst of UK’s Covid-19 lockdowns, 22 participants across the UK provided 2,054 unique patient-days of data in phase 2 of the AAMOS-00 Study.
The full data dictionary is available via the DataShare website
The sci_data_AAMOS-00_start_up.R file can be used to get started this dataset, it includes joining data tables, data wrangling, and an example binary classification problem.
The code is used to produce the plots and results presented in the "Technical Validation" of our Scientific Data paper.
Begin by downloading all 12 files from the DataShare page. Then running all the lines of code in sci_data_AAMOS-00_start_up.R will produce all the plots and results. The sci_data_AAMOS-00_start_up.R file contains 4 sections (Import data, Data processing, Binary classifier example, Technical validation (asthma symptoms)), each section depends on the previous section.
Data is licensed under the Creative Commons Attribution 4.0 International Public License.
The starter code is licensed under the MIT License.
If you would like to collaborate or request data from the AAMOS-00 study, please direct your correpondence to Kevin C.H. Tsang (kevin.tsang@ed.ac.uk) or Syed Ahmar Shah (ahmar.shah@ed.ac.uk)