Context

Environmental sensor telemetry data, detailed in the blog post, Getting Started with IoT Analytics on AWS, published on Towards Data Science.

Content

The data was generated from a series of three identical, custom-built, breadboard-based sensor arrays. Each array was connected to a Raspberry Pi devices. Each of the three IoT devices was placed in a physical location with varied environmental conditions.

| device      | environmental conditions         |
|-------------------|------------------------------------------|
| 00:0f:00:70:91:0a | stable conditions, cooler and more humid |
| 1c:bf:ce:15:ec:4d | highly variable temperature and humidity |
| b8:27:eb:bf:9d:51 | stable conditions, warmer and dryer   |

Each IoT device collected a total of seven different readings from the four sensors on a regular interval. Sensor readings include temperature, humidity, carbon monoxide (CO), liquid petroleum gas (LPG), smoke, light, and motion. The data spans the period from 07/12/2020 00:00:00 UTC – 07/19/2020 23:59:59 UTC. There is a total of 405,184 rows of data.

The sensor readings, along with a unique device ID and timestamp, were published as a single message, using the ISO standard Message Queuing Telemetry Transport (MQTT) network protocol. Below is an example of an MQTT message payload.

{
 "data": {
  "co": 0.006104480269226063,
  "humidity": 55.099998474121094,
  "light": true,
  "lpg": 0.008895956948783413,
  "motion": false,
  "smoke": 0.023978358312270912,
  "temp": 31.799999237060547
 },
 "device_id": "6e:81:c9:d4:9e:58",
 "ts": 1594419195.292461
}

Columns

There are nine columns in the dataset.

| column  | description     | units   |
|----------|----------------------|------------|
| ts    | timestamp of event  | epoch   |
| device  | unique device name  | string   |
| co    | carbon monoxide   | ppm (%)  |
| humidity | humidity       | percentage |
| light  | light detected?   | boolean  |
| lpg   | liquid petroleum gas | ppm (%)  |
| motion  | motion detected?   | boolean  |
| smoke  | smoke        | ppm (%)  |
| temp   | temperature     | Fahrenheit |

Detail Image of Sensors

https://programmaticponderings.files.wordpress.com/2020/07/rasppi_detail-1.jpg" alt="Raspberry Pi Sensor Arrays">

The Sentiment140 Twitter sentiment dataset analyzed in this study is publicly available and can be downloaded directly from: (https://nyc3.digitaloceanspaces.com/ml-files-distro/v1/investigating-sentiment-analysis/data/training.1600000.processed.noemoticon.csv.zip). The Amazon Customer Reviews dataset is publicly available via the AWS Registry of Open Data at: (https://registry.opendata.aws/amazon-reviews/). (ZIP)