Looks pretty solid, congrats!

Didn't you get the memo? We're all on the Lovable band wagon now

There’s a big difference between the origin of a game idea and improving an existing idea based on feedback. I’m not saying you should blindly follow player feedback.

You can focus on your vision but utilizing feedback to improve player retention at the same time

I remember downloading (completely legal of course🙄) a song and trying to put it on a floppy disk!

looks dope!

Thanks for reading and making it until the end! Any suggestions on which game I should do next?

Also, just for the heck of it, I used AI to create a podcast out of all the reviews and this actually pretty scary. It resulted in a 15-minute podcast with two hosts going back and forth about Timberborn. I’ll let you experience it yourself here:

soundcloud.com/tim-8703501...

I've also queried the dataset with specific questions about the players' reception of the update, but for the sake of the length of this post I'll refer to the original blog post here where all the info is:

www.resonate.gg/post/timber...

People also seemed to like the Golem/bots update in September of 2022, which caused sentiment to rise sharply. It dropped again after changes were made to storage (storing only one item in warehouse instead of multiple).

So in short: Sentiment was positive all around, except the Badwater update didn’t perform really well, dropping sentiment to its lowest point. Wonder of Water performed really well, making sentiment climb higher than it has ever been.

When we add reference lines for the recent update to Timberborn, things are becoming a bit clearer:

There’s still a lot of noise in the data though, even the moving average doesn’t really help. So here it is aggregated by month:

Keep in mind that sentiment can range from -1 (very negative) and +1 (very positive). Average sentiment for the entire time period is 0.47, which is very positive, which I think also reflects what I seem to read on Reddit and what is reflected in the Steam score.

Blue line is average sentiment on a daily basis. Red is a moving average, painting a clearer picture. Sentiment was positive and around 0.6, except for one particularly bad in May of 2022. From what I found this was just bad luck with a couple of negative reviews on that day.

Having only English reviews I ran sentiment analysis over every individual review with machine learning. There's a small error rate and it doesn't always correspond to the recommended setting of the review, but on an aggregate level it paints a picture. Sample period is from release until Feb 2025.

Of all 28,937 reviews I extracted from Steam only 61,9% were in English. To get a complete picture and to cover the entire player base, I had to translate them, which I did by running a local translation model. Here's the language distribution of the reviews:

Sounds like a great idea for a new post in which you thank your entire family, friends and colleagues for this opportunity!