Seaside Cloudworks

A recap of the previous session

Last time, I created a new smartphone app, set up a data flow to retrieve information from Google HealthConnect, and wrote it to a Google Spreadsheet via Google Apps Script. This is the red route shown below.

However, the data currently available is shown in the table below; the step count is significantly higher than the actual number, and other fields are left blank. This is likely because the collected data was skewed toward “steps” or included extra numbers.

Improvements to Health DX Tools

In this update, we will resolve the aforementioned issues and ensure that step count, weight, body fat percentage, and sleep duration are correctly collected from HealthConnect and recorded in the spreadsheet. We will also implement the following improvements:

• Identifying and Resolving Issues with Measurement Data
• UI Development

Ultimately, as shown in the figure below, I had been manually entering data into the two apps on the left (Appsheet and Google Fit)—specifically, I had been manually entering every single field in Appsheet. I also entered my weight and body fat percentage separately into Google Fit and Fitbit, using each app for its respective purpose. To address this, I aim to streamline the process by consolidating data retrieval from HealthConnect and inputting it directly within the app, thereby reducing the burden of data entry.

Step 1: Identifying the Cause of the Problem

Problem: I was aggregating data from multiple data sources (Origins).

The issue with the previous data—where the step count exceeded my actual experience—was that it had combined data from multiple sources (Origins).

In my setup, step counts are tracked by both my smartwatch (Fitbit) and my smartphone, and in this case, the data is recorded on HealthConnect separately by source (Origin). The previous result was inflated compared to reality because it had summed the data from multiple sources (Origins).

When I connected my smartphone to Android Studio and used the debugging feature, I discovered that there are three types of data sources (Origins): Fitbit, android.apps.fitness, and android. I assume that the two starting with “android” both represent the same value and correspond to HealthConnect and the Android OS, respectively.

If you have smartwatches, pedometers, or other devices with write permissions for HealthConnect, the number of data sources (Origins) increases, so you’ll need to select one as needed. In this case, I’d like to use Fitbit as the priority device, but there was another issue. Please take a look at the data from a different day below.

The Fitbit reading for this day looks quite low, but that’s actually because there were times when my smartwatch ran out of battery and couldn’t track my activity. Since my device needs to be charged every few days, if I forget to charge it, the battery dies during the day and it stops tracking. Also, since I leave it on the charger while it’s charging, it’s unable to track during that time as well. During those periods, I tend to rely on the step count from my smartphone instead.

However, even these 6,108 steps feel significantly lower than I actually felt. In fact, when I walk or run on a treadmill, I tend to set my smartphone aside, so some steps may not be counted—that’s another issue I need to address.

Solution: Set the priority for Origin and provide a manual adjustment screen.

We will address these issues in accordance with the following specifications.

• (Specification 1) Generally, among multiple Origins, the Origin with the highest daily value is selected.
• (Specification 2) This app also allows direct data entry into HealthConnect to add data to a new Origin. *This is intended to correct data when the data from any of the Origins appears abnormal.

As for other issues, we will first clarify the specifications and proceed with development by adding them to the backlog to ensure nothing is overlooked.

Tips: Have ChatGPT manage your chat history

I use ChatGPT to manage the backlog of tasks that need to be addressed. In my case, I use prompts like the one below to manage the backlog and check on its status.

Add to Backlog

Please manage this item as a backlog.
・Among multiple Origins, select the Origin that recorded the highest daily value.
・Enable direct data entry into HealthConnect within this app and add data to a new Origin.

Check the backlog status

Please provide a breakdown of the current backlog by status.

Consider the next steps

Which task should I tackle next?

The prompt I’m still using is designed to sound like a simple conversation, but it’s generally performing as expected.

Step 2: How to Import Values

In the first stage, we confirmed that there is a value for “Steps” for each Origin, so we will review the values for the other items and consider how to import them.

Problem: Characteristics by Category and Issues to Be Addressed

Item	Origin	Session Tracking	Issue
Steps	Fitbit, Device	Yes	Multiple sources, manual entry
Sleep times	Fitbit	Yes	Daily boundaries, multiple sessions
Weight	Google Fit (manual entry)	No	Manual entry
Body Fat%	Google Fit (manual entry)	No	Manual entry, decimal input

As with the step count mentioned earlier, when reviewing the debug logs, I noticed the characteristics described above, and it was particularly necessary to address the “session status” and the issues outlined in the assignment. Below are some additional notes regarding the assignment.

• Session Status: Step count and sleep duration are recorded not only as values but also with start and end times (e.g., 07:00–09:00); these are referred to as sessions. Throughout the day, there are frequent breaks—such as walking, resting, sleeping, waking up (e.g., to use the restroom), or taking a nap. You need to combine the values from multiple sessions. Weight and body fat can be recorded multiple times a day by specifying the date and time, but for the purpose of compiling daily results, it is simple because you only need to collect the latest values.
• Multiple Origins: Since values are recorded for each data source (Origin) mentioned in the section on step count above, you need to select which values to use.
• Manual Entry: In my setup, I need a way to manually enter weight and body fat percentage—which aren’t recorded automatically—as well as steps and sleep duration, since these can’t be recorded when the battery is dead or the device is charging.
• Daily Boundaries: Where should sleep duration be divided? For example, how should we handle a scenario where someone goes to bed at 10:00 PM and wakes up at 6:00 AM? Should we count it as 8 hours of sleep for the previous day, or split it at midnight to record 2 hours and 6 hours on separate days?
• Decimal Input: While weight and body fat percentage can be manually entered in Google Fit and recorded in HealthConnect, body fat percentage can only be entered as an integer (e.g., 22%). Since I wanted to enter values with decimals (e.g., 21.8%), I need a way to input them.

Solution: Consider how to import the data based on the characteristics of each value

We have also included specifications 1 and 2 mentioned earlier.

• (Specification 1) Generally, among multiple Origins, the Origin with the highest daily value is selected.
• (Specification 2) This app also allows direct data entry into HealthConnect, adding data to a new Origin.
• (Specification 3) Step count and sleep data are summed across multiple sessions.
• (Specification 4) Sleep duration is calculated by setting a cutoff time to aggregate values from the previous day. Example: 6:00 cutoff, etc.
• (Specification 5) When entering data manually, allow the input of body fat percentage values that include decimals.
• (Specification 6) To prioritize manually entered values, review the priority of Origins.
• (Specification 7) When entering sleep and step counts, do not create session information or upload data to HealthConnect; instead, manage the data as temporary data on the device.

Regarding Specification 7, since creating session-based data would be too complex, and since we only need to know how many steps were taken and how many hours of sleep were recorded each day, we have decided not to create session data.

The actual development process involved reviewing debug logs in Android Studio to understand the specifications, generating source code using ChatGPT, modifying the source code in Android Studio, rebuilding the app, connecting it to a smartphone, and debugging—repeating this cycle to gradually improve the code. Eventually, I managed to get it to the point where values are recorded in a spreadsheet, as shown below.

I found that the approach of managing the backlog with ChatGPT and building it incrementally helped me understand the current state of affairs and clarify what needed to be done, making the process much smoother. However, I’ve noticed that as the volume of information increases—such as pasting source code or logs into the conversation or having ChatGPT generate code—the dialogue becomes increasingly sluggish, which is a challenge. I’d like to try using modern AI agents in the future to see if there’s a more efficient way to handle this.

Step 3: UI and GAS Development

The user interface (UI) was developed primarily with the following features. It is the result of continuous improvements made by adding necessary features as needed, starting with a version created for functional testing.

Screen	Features
Initial Screen	GAS Communication Settings and Operation Check
Main Screen	Daily main screen, synchronization feature, access to input and settings
Input Screen	Function for entering weight, body fat percentage, step count, and sleep duration as needed
Setting Screen	GAS communication settings, sleep duration cutoff times, etc.

Since I enter my weight and body fat percentage every day, all of these values are adjusted figures. As for the step count on March 12, my Fitbit ran out of battery, and since I had left my phone behind, I’ve entered an approximate value.

Finally

Using the smartphone app I developed, Hc2Spread (HealthConnect to Spreadsheet), I retrieve step counts and sleep duration data from Fitbit and Google Fit. For weight and body fat percentage data, which aren’t imported automatically, I can now enter them into HealthConnect directly from this app. If the values are incorrect, I can now edit them in the app and update the data in the spreadsheet via Google Apps Script (GAS).

Ideally, I would have liked to eliminate the app on Appsheet, but since it excels at visualizing data through graphs, I’ve decided to keep using it while reducing the number of input fields. It’s hard to say for sure whether this constitutes an improvement, but I’m quite satisfied that I’ve been able to use generative AI to connect to HealthConnect and store the data in a spreadsheet for easier use.

Going forward, I’d like to explore “automation” and other approaches to further enhance the effectiveness of this system.

Health DX Progress Confirmation

This time, we will again identify the progress and challenges of Health DX tool users. Approximately one month has passed since the last session.

• Weight: Continuing to decrease at a pace of 1kg per month. Progress is relatively steady.
• Body Fat Percentage: The body composition monitor was switched midway, causing the readings to change.
• Steps: Around 8,000 steps, with more days falling slightly below that.
• Sleep: Continually getting less than 5 hours most nights, which is on the low side.

Last time, target values were set for each graph, and a moving average was added to the weight graph. This makes the targets easier to understand as benchmarks, and the moving average is also helping shift my mindset away from getting overly excited or disappointed by daily fluctuations.

Advantages and Challenges as a Health DX Tool

I will summarize the strengths and challenges based on my actual experience, including what was implemented in the previous initiative.

• Positive Points
  • The improved usability of entering meals via buttons reduces the burden of daily input.
  • The target values on each graph provide clear daily benchmarks.
  • The addition of a weighted average for weight makes it easier to avoid getting overly excited or disappointed by daily fluctuations.
• Challenges
  • Re-entering values already available on the smartphone (such as step count) is burdensome.

This time, I’ll continue making improvements in consultation with my trainer, but I’ll omit diet advice from Health DX. This is because progress has been relatively stable, and my trainer recommends continuing as is. This time, I’ll primarily focus on improving the tools.

Improving Health DX Tools

This time, we’ll consider improvements to “re-entering values already present on the smartphone.” For example, step counts and sleep duration are obtained by smartwatches or the smartphone itself, and on Android smartphones, they are accumulated in Google Health Connect. The functional structure is shown in the diagram.

I check the values in the Fitbit or Google Fit app, transcribe them into Appsheet, and ultimately record them in Google Spreadsheet via Appsheet. I want to reduce this transcription work. Specifically, I plan to develop a new app on my smartphone, as shown in the diagram below. This app will retrieve data from Health Connect and register the values in Spreadsheet via Google Apps Script.

This countermeasure involves establishing this countermeasure route. However, establishing this route presented significant challenges this time, so we have summarized the issues and countermeasures we encountered while leveraging generative AI to drive improvements.

The First Wall: Understanding build and dependency structures

This time, it was my first time developing an Android app (using Android Studio), my first time using Kotlin, and my first time using Gradle—a whole lot of firsts. While I’m using ChatGPT to help with the development, my lack of foundational knowledge was so severe that it took a considerable amount of time just to get the build working. I incorporated the generated source code and dependencies for accessing Health Connect into the simplest possible screen app in Android Studio, but I got stuck because it wouldn’t compile. The main issues were as follows:

• Specifying SDK versions (compileSdk, targetSdk, minsdk)
• Minimizing dependencies
• Unsynchronized state due to not performing Gradle Sync

1.Specifying SDK Versions (compileSdk, targetSdk, minsdk)

The main source is MainActivity.kt, dependencies are specified in build.gradle.kts, and we build using Android Studio. We pass build errors to ChatGPT and iterate on improvements. The above two issues were resolved by specifying three SDK versions.

• compilesdk = 36 # SDK version used during compilation
• targetsdk = 34 # SDK version at runtime; slightly lower than compilesdk to resolve build errors
• minsdk = 26 # Minimum guaranteed SDK version; specifies a version where Health Connect is available

Specifying the exact version was crucial. However, I didn’t thoroughly investigate the rationale behind each value. I had ChatGPT figure it out, and since the build wouldn’t pass, I ended up lowering the target SDK.

2.Streamline dependencies to the minimal configuration

This is also part of my trial and error with ChatGPT. It seems there was a conflict within the implementation that was included in the initially generated source code. When I tried passing the source to a different chat again, the results were narrowed down to the following two:

    implementation("androidx.health.connect:connect-client:1.1.0")
    implementation("com.squareup.okhttp3:okhttp:4.12.0")

The first line is the Health Connect library, and the second line is the library used to call Google Apps Script (GAS).

3.Unsynchronized State Due to Not Performing Gradle Sync

I hadn’t realized this due to my inexperience, but I assumed that updating the Gradle definitions would automatically keep dependencies up to date and load the necessary libraries. However, the “Gradle sync” operation was required.

In Android Studio, you can perform “Gradle sync” using the following screen operations:

• Click the elephant icon labeled “Sync Project with Gradle Files” in the upper-right corner of the screen
• or press “CTRL + SHIFT + O”

The Second Wall: Lack of Understanding of the Execution Environment and Permissions Model

This was partly due to my inexperience and reliance on generative AI, which led to gaps in my knowledge. The code generated by the AI didn’t run as-is; I had to modify it to resolve issues before it worked. I thought I understood this, but I still struggled quite a bit.

Specifically, when determining if Health Connect was active, I failed to notice that the meaning of the numbers differed depending on the SDK version. For a while, I mistakenly believed Health Connect wasn’t usable. Ultimately, the Health Connect check was implemented as follows:

val status = HealthConnectClient.getSdkStatus(this, HC_PROVIDER)
val statusLabel = when (status) {
    HealthConnectClient.SDK_AVAILABLE -> "SDK_AVAILABLE"
    HealthConnectClient.SDK_UNAVAILABLE -> "SDK_UNAVAILABLE"
    HealthConnectClient.SDK_UNAVAILABLE_PROVIDER_UPDATE_REQUIRED -> "SDK_UNAVAILABLE_PROVIDER_UPDATE_REQUIRED"
    else -> "UNKNOWN($status)"
}
Log.d("HC", "Health Connect SDK status = $statusLabel ($status)")

This confirms that Health Connect is now available, allowing us to proceed.

The Third Wall: Asynchronous Processing and Network Design

There are many things you won’t notice unless you know about them. For example, there’s a rule stating that “Communication must not be an extension of button operations.” The initial source code generated by ChatGPT also treated it as communication extending from button operations. However, the issues mentioned above were pointed out, so I checked the countermeasures with ChatGPT and made improvements.

Specifically, I implemented the following three countermeasures:

• Change post-processing to a suspend function
• Wrap with withContext(Dispatchers.IO)
• Add Coroutine import

It’s easy to write about, but I couldn’t solve it myself—it wouldn’t have improved without ChatGPT. I wish it had been designed to avoid problems from the start, but I’m satisfied that it’s resolved now.

Add Google Apps Script and update the spreadsheet

I also generated the source using AI and registered it. When I pointed out that Spreadsheet didn’t recognize the processing, they added it, and it executed surprisingly smoothly. The source is simple, with only about 20 to 30 lines.

Upon execution, values were recorded in column A (date) and column B (steps). Looking at column F, you can see that it was run twice over about four days. Each time, it aggregates and records data from the past three days.

Finally

I successfully connected to Health Connect and registered the data in Google Spreadsheets, but this data is actually incorrect. The numbers are significantly larger than the actual values, and I haven’t been able to capture weight, body fat percentage, or sleep duration, so further improvements are needed. I plan to address these issues in the next opportunity.

For now, I’ll consider it a success that I managed to punch holes in the series of processes and wrap things up here.

Health DX Progress Confirmation

This is a continuation of our discussion about building our own employee health management system. It’s been about two weeks since we migrated from spreadsheets to Appsheet, and the data has continued to grow. As before, we’ll further review both the weight loss initiative as part of our health DX efforts and the underlying mechanisms.

• Weight: Overall decreasing trend, averaging about 1kg less.
• Body Fat Percentage: Also decreased by about 1%.
• Steps: Averaging over 9,000 steps.
• Sleep: Mostly under 5 hours, on the low side.

I feel like I’m starting to see some results, but I’ll continue working on the following: addressing operational challenges, reviewing the status of my health DX with my trainer, and making improvements.

Advantages and Challenges as a Health DX Tool

Here’s a summary of the pros and cons we experienced after implementing AppSheet in our previous initiative.

• Pros
  • Input can now be done on smartphones, making it easier to enter data anywhere.
  • Visibility improved by using graphs on the dashboard.
• Cons
  • Make it clearer by including target values to aim for.
  • Weight fluctuates a lot, so make trends easier to understand.
  • Since values are entered manually, we want to make input easier.

First, we’ll proceed by discussing the current situation with the trainer.

Current Status Review and Future Action Planning with the Trainer (ChatGPT)

We will provide the current data to the trainer to review the current figures and discuss future actions. The following are ongoing initiatives from the previous session.

Current Situation Analysis and Countermeasures

As before, we provided the data to the trainer to confirm the current status. The trainer gave us the harsh opinion that “it’s still too early to say we’re seeing a downward trend.” However, since the numbers do show some reduction, they suggested we try the following measures.

Regarding sleep, I’ll focus on effort, and for meal tracking, I’ll explore improvements within the tool itself.

Health DX Tool Improvements

We’ll address the issues mentioned at the beginning and the trainer’s feedback on meal tracking within the tool. This time, we’ll focus on the following four points:

• Measure 1: Add Target Values: Include target values to make goals clearer.
• Measure 2: Add Moving Averages: Weight fluctuates a lot, so moving averages will help show trends more clearly.
• Measure 3: Improve Input Method: Currently manual entry; make input easier.
• Measure 4: Add Meal Logging: Add a dedicated screen and data for meal records.

Measure 1: Add Target Values

While numbers are now visible on dashboards and graphs, without targets, it’s unclear how much effort is needed. We’ll set target values for each item and display them on the graphs.

1.Adding Target Values

Add a “Target Values” sheet to the spreadsheet where data is recorded and set target values for each item.

2.Link the data recording sheet with the target value sheet in Appsheet

Use the Lookup function to add a virtual column and set up cross-sheet references using the Lookup function.

Set up the target value sheet to load in Appsheet as before.

Next, we’ll add a link between the two sheets. Since the chart references data from Sheet1, open Sheet1 and add a Virtual Column (VC) with a Lookup function for the link. First, add the VC.

Click the Add Virtual Column button to add an item. Set the formula as follows. This Lookup function references another sheet and sets the value in this VC.

Adding an item will place it at the bottom of the table as shown below.

Next, configure the graph settings. Open the UX Weight section, click the Add button under Chart columns, and add the “Weight (Goal)” column you added earlier. After adding it, the display will appear as shown below. The graph colors are also set individually. The target value is uniformly displayed in red.

Adding the target weight is now complete. Similarly, add the VC and graph for each item. Once all additions are finished, it will look like this. For the target weight, while the ultimate goal is 60kg, it feels too distant, so I’ve set the immediate target to 67kg to keep it within an achievable range. Everything else remains unchanged.

This should make it easier to identify your target compared to just looking at actual values, right?

Countermeasure 2: Adding a Moving Average

The trainer said, “Since a person’s weight fluctuates by about ±0.8kg daily due to body water levels, it’s better to take a moving average over about a week to see the trend.” So, first, we’ll add a moving average to the weight section. The method for adding it is the same as above: add a VC and include the calculation formula.

Definition of this moving average: Calculate the average of the weight data for the week including the current day (from 6 days prior to the current day), and use this as the moving average value for the current day.

The VC settings are as follows:

Adding this item to the UX weight will enable the moving average.

A blue broken line has been added to the graph on the right. While the actual weight values fluctuate wildly, you can see that the moving average shows a smoother progression. It should now be easier to see that the weight has decreased slightly compared to the beginning.

Countermeasure 3: Improve Input Methods

Currently, input is limited to numerical fields and text notes, with text input or numerical adjustments via the plus/minus icons on the far right. We will make input methods more user-friendly by selecting them based on the data type.

The sleep duration input method will be changed to a slider input. To display the slider, select Range and set it to a range of 3 to 10 hours in 0.2-hour increments.

Try adjusting each setting to experiment with different options. In the next meal log section, we’ll test the button input method.

Countermeasure 4: Adding Meal Logs

Since meal preferences vary, please customize the following as a reference example for using the tool. The trainer mentioned “fried foods no more than twice a week,” so we’ll add a field for whether fried foods were consumed.

Set up the management fields as follows:

• Breakfast: Enter common breakfast items via buttons
• Breakfast Notes: Enter other items as text
• Add similar items for Lunch, Dinner, and Snacks
• Fried Foods: Indicate whether fried foods were consumed that day

First, add data columns to the spreadsheet. The newly added columns are highlighted in green for clarity. You’ll notice a column labeled S1 at the top; this is included for page control on the input screen. It might be possible to omit it, but for now, it’s included.

Below are the methods for switching input pages, entering meal buttons, and indicating fried food presence.

1.Switching Input Pages

Configure page switching using the S1 item added above. Set S1 as follows: Set Type to Show and Category to Page_Header. This will display “Next” next to the Save button on the right, and the meal item added above will appear on the next page.

When switching pages, if a message in red text appears indicating that an entry is required, uncheck the “Require?” box next to the data field in the list. Currently, it’s best to leave all boxes unchecked except for the date field.

2.Adding Meal Input Buttons

For example, if you regularly eat certain items every morning like eggs, bread, and milk, adding these as buttons allows you to input them all at once. Items you don’t eat daily should be listed in the text input field.

Set the Type to EnumList, press the Add button for Values, and enter items like eggs, bread, and milk as shown in the figure below.

Once saved, you can input items by clicking buttons as shown in the lower right screen, then add any missing items in text. If the items used change, we’ll review the button contents and customize them for easier use.

3.Whether you ate fried food

If you eat fried food with every meal, it’s better to add a fried food button to the top buttons. However, if you want to manage it only once a day or so, keep it as a separate item like this time. It might also be clearer for your trainer to see if they’re separated.

Since this item asks whether you ate fried food with a Yes or No answer, it shouldn’t allow multiple selections like the top buttons. It must be one or the other.

Simply set the Category to Yes/No. This automatically generates a button where only one option can be pressed. You can also change the button text in the settings.

Next, add lunch, dinner, and snacks in the same way to complete this improvement.

Finally

This time, we’ve further reviewed visualization and input methods, making daily use easier. Comparing against target values should also boost motivation.

However, as the trainer pointed out, it’s important not to get too caught up in daily fluctuations. It’s better to accumulate data weekly or over a period and then review it together with the trainer. Currently, we’re conducting weekly reviews.

I’d like to continue operating in this state for a while.

About two more weeks have passed since last time, and the data has accumulated. I’d like to analyze the accumulated data and consider improvements to the current functionality.
First, I’ll review the data.

Overview from the Data

• Weight: Remained mostly flat. It appeared to decrease slightly for a while but has since returned to its previous level.
• Body Fat Percentage: Decreased by 1%
• Steps: Maintained around 8,000 steps per day
• Sleep Duration: No significant change

Later, I’ll consider what to do next with this data to achieve our goals.

Challenges as a Health DX Tool

The biggest issue is the significant operational burden required to maintain daily records. Since the spreadsheet must be opened on a smartphone, the following advantages and disadvantages arise:

• Advantages
  • Easy to add items and graphs (feels familiar since it works like Excel)
  • Accessible from various devices like PCs and smartphones
• Disadvantages
  • Reaching the input screen on smartphones requires a slightly cumbersome process
  • All text input must be done manually

We will also work on improving the tools later.

Discussing current analysis and future weight loss plans with the trainer (ChatGPT)

First, we will conduct a “current situation analysis” with the trainer, followed by “consideration of improvements.”

1.Current Situation Analysis

Provide the trainer (ChatGPT) with the data as a table and give the following prompt to ensure it correctly recognizes the data.

I’ll attach the data recorded in the app I created for a while. Can you tell what each item represents?

I think I could have written it a bit more specifically, but for now, I tend to convey things in a natural way. It’s just that I can’t write long, structured prompts. The trainer provided the following interpretations for each item. It seems I understood correctly.

※The response contains more detailed information, but this is an excerpt.

2.Improvement Considerations

My diet doesn’t seem to be progressing much as it is now, so I’ll ask the trainer for their thoughts.

Regarding my weight trend, will I be able to maintain a pace that keeps me below 60kg by the end of September 2026?

After various explanations, the conclusion is as follows: Achieving the goal is still possible if you can lose 0.98 kg per month for the remaining period.

Various countermeasures were suggested, but the summary is to continue as is for a while longer.
The advice was to consider increasing sleep time by even 30 minutes. For now, I’ll continue maintaining the current status.

Currently, I’m focusing on the following points. To those who exercise regularly, this amount might seem negligible.

• Exercise: Walk 8,000 steps per day, morning and evening stretching, and light strength training (20 sit-ups and 10 push-ups without equipment).
• Diet: Veggie-first (salad first at every meal), reduce white rice only at dinner (no reduction in noodles, etc.), add protein to breakfast.

Since I believe muscle mass decreases first when weight drops without much exercise, I supplement my protein intake with protein powder. I also consulted my trainer about my diet and introduced this. For now, I plan to trust my trainer and stick with this approach.

Improving Health DX Tools

We will maintain the previously mentioned advantages—“ease of modification (adding items/graphs)” and “operability across multiple devices”—while considering improvements to the following disadvantages.

• Disadvantages
  • The steps to reach the input screen on a smartphone are a bit cumbersome.
  • All text input must be done manually.

I consulted ChatGPT about this too. I had assumed we’d end up creating a smartphone app, but it recommended using Google AppSheets for no-code development. It also mentioned that using this tool would allow us to reuse the spreadsheets we’d already been accumulating, which is incredibly convenient. Even if we end up not liking it, we can easily revert back to our current workflow. This seems like a safe approach to take.

Google AppSheet

https://about.appsheet.com/home

1.Getting Started with AppSheet

Sign in to the above site to enable AppSheet. After signing in, you will see a screen like the one below.

2.Create a data display using the currently used spreadsheet

Select Create > App > Start with existing data in the upper-left corner of the screen, specify the spreadsheet you’ve been using, and click Submit. The following three images illustrate this process.

Select Google Sheets.

Select the target file and submit.

Okay. Done. The tall vertical view on the right side of the screen shows how it will appear on a smartphone. You can also test operations on this screen. Clicking the blue + (plus) icon allows you to add data. Clicking a daily tile lets you update that day’s data in the items around the center of the screen, which correspond to the original spreadsheet columns.

3.Add a visualization screen

You can add visualizations like graphs from the UX menu on the left side of the screen. While I’ve combined weight, body fat percentage, step count, and sleep duration into a single line graph, various other representations are possible, as shown in the center of the screen. You can choose based on your purpose.

4.Create a dashboard

Similarly, you can also add dashboards from UX. This time, I added the four graphs created above to display them as a series. However, on smartphones, you’ll view them with vertical scrolling, as shown on the right in the figure.

Switching to tablet view allows you to see everything grouped together like this diagram. Thanks to its responsive design, it works seamlessly across various devices like PCs, tablets, and smartphones.

5.Make it available for use on iPhone and Android devices

For everyday use, we’ll make it available as an app. Both iPhone and Android have an app called “AppSheet,” and using this app allows you to access the app created above.

Finally

This resolves the issues previously listed as drawbacks, making daily data entry easier and more sustainable. Consequently, we believe it will be easier to achieve improvements in employee health.

We plan to continue using the system, resolving issues and pursuing further improvements. Moving forward, we will use the smartphone app for data entry and updates, while using PCs for regular analysis and data transfer.

We have started using AppSheet for meal logging, so next we want to focus on refining the input screen.