At Angaza, we’re excited about the extra value we provide our customers and their clients by leveraging the power of data. One particular area of focus for our product and data science teams is delivering more valuable insights from smart device data.
Devices that contain embedded sensors regularly sampling system parameters and transmitting information for remote monitoring are often referred to as “smart devices” or the “things” in the Internet of Things (IoT). IoT systems generate enormous amounts of data essential to the operations of logistically complicated or geographically dispersed businesses. Distributors of last-mile products are perfectly positioned to benefit from the insights and operational efficiencies IoT based analyses can provide.
IoT by Angaza is a fully integrated turn-key solution combining embedded IoT technology, robust connectivity, software infrastructure, and powerful data products. Angaza works directly with device manufacturers to enable reliable device-level telemetry through intelligent monitoring and robust connectivity. When our customers deploy smart devices in the field, Angaza’s software infrastructure manages device communication, data transmission, and storage. This data is then transformed into powerful data products through our proprietary data analysis and algorithms, which are available to our customers on the Angaza Hub.
These smart devices have several sensors onboard that measure a range of electrical parameters at regular intervals, such as every hour, and can transmit on demand or on a personalized schedule. IoT by Angaza communicates with devices over the widespread 2G GSM (Global System for Mobile communications) network. Consequently, we receive data in near real time, allowing Angaza customers to remotely monitor the system performance and understand how the device is being used.
Smart Device Data
We work with our manufacturing partners to ensure reliable device-level telemetry which monitors and reports valuable system information. For solar devices, that information includes the four essential variables related to system performance:
- Solar Panel Voltage: This measurement indicates how much solar power is available.
- Solar Panel Current: This measurement indicates how much solar power is being used to charge the battery.
- Battery Voltage: This measurement indicates how much energy has been stored and is available to be used.
- Battery Discharge Current: This measurement indicates how much power the client is using.
This stream of sensor data is typically referred to as usage data. In addition to usage data, we also retrieve information about the location of a device from the GSM network when the device communicates with Angaza. At present these datasets produce millions of data points every week! Our customers can view the usage and location data for individual smart devices via the Angaza Hub.
Given the wealth of data being collected, we continuously ask ourselves, How can we provide value to our customers from this dataset? The answer is that there are many ways; including remote fault detection, on-demand location services as well as aggregate insights about client usage trends.
Portfolio Insights From Location Data
With non-smart devices, once a client has purchased a system, our customers typically have limited knowledge about where that device is installed. Other than contacting the client to confirm the location, it is often assumed that it is at the address of the purchasing client. However, with IoT by Angaza, our customers and their clients are kept up-to-date about the location of their smart devices.
For a portfolio of smart devices, this information becomes a fantastic operational tool. Figure 1 shows a heat map of device density in Uganda, where red indicates more devices and blue indicates fewer devices. Device locations have been simulated in order to protect the confidentiality of our customers and their clients.
Figure 1: Simulated heatmap of smart device density in Uganda
This type of data can be used for optimizing post-sales support teams or for evaluating the success of regional marketing campaigns. For example, it may make sense to focus technical support in areas of high product density in order to better resolve faults, if or when they arise.
Several of our customers have also expressed that their clients are concerned about theft. In this case, location data can act as an excellent deterrent and be used as a tool to locate the device if theft ever does become a reality.
Case Studies in Client Usage
Two early adopters of smart devices are Angaza customers Mwezi Limited and Sosai Renewable Energy Company.
Thanks to IoT by Angaza, Mwezi and Sosai can now start to answer the question, How are my clients using their device? Behavioral analytics can play an important role in marketing campaigns, strategic decision making, and after-sales support. One way to approach this question is to look at what we call the ‘Daily Energy Consumption Profile,’ which describes how a device is used over an average day. Consumption is the amount of energy used by the client in a given hour of the day and is measured in Watt-hours (Wh).
Figure 2 shows the typical daily consumption profile for a subset of devices (light blue lines) and the average is highlighted by the solid blue line. We can see that although there is significant variation across clients, on average there is a strong peak between 8 pm and 9 pm. One interesting fact to note is that the smart devices analyzed here supply power to televisions among other accessories, such as lights. Therefore one theory we have is that this strong peak corresponds to recreational TV use. Equally, there is a smaller peak between 2 and 3 pm.
Figure 2: Daily Energy Consumption Profiles for a subset of devices and their average.
Aggregate data used with customer permission.
This information can be used for marketing campaigns and after-sales support, among others. Given that clients on average use their devices during the evening, a marketing campaign focused on evening activities can be developed. Additionally, maintenance teams can use this information to identify the best times to service devices; it seems early morning would be best as this is when clients typically use their devices the least.
With regards to after-sales support, typically our customers only become aware of potential device faults if and when the client notices abnormal behavior and contacts the support team. With IoT by Angaza, our customers are able to proactively detect and resolve faults over faster timelines.
Figure 3: Daily Solar Power Profiles and their average. Aggregate data used with customer permission.
As an example, solar panel voltage is indicative of how much solar power is available to charge the device’s battery; the higher the solar panel voltage, the more power is available. This is shown in Figure 3, where we can see the solar panel voltage increasing during sunlight hours. Such data can be used to determine whether the system is operating correctly; given there is sufficient solar power and the battery level is low, the battery should begin to charge automatically. If this is not observed, then there may be a fault with the unit.
Our data science team has developed a set of proprietary fault detection algorithms that automatically identify this and other potential fault scenarios, preemptively alerting a distributor about affected devices through the ‘Unit Health Alert’ feature on the Angaza Hub. Fault scenarios are classified into four alert categories; solar, battery, usage, and data for additional visibility into the type of fault that triggered the alert. For our customers, these unit health alerts can provide significant value to their post-sales support teams and give their clients an improved product experience.
IoT by Angaza Hub Features
Usage and location data are just a few of the ways that IoT by Angaza brings added benefits to our customers. The Angaza product and data science teams are committed to maximizing the value of smart devices by creating features that allow our customers to have unprecedented insights into the real-time states of the devices and to provide first-class after-sales support to their clients.