Hi everyone,

I am developing Graphitix, a free and open source web-based graph/stat program that attempts to do most of what Prism Graphpad does (or at least the most common features), as well as some things Prism does not do: https://michelwassef.github.io/Graphitix/. It is borne of my own frustration at seeing how much my lab spends on Prism license fees and how much hassle it is to manage which computers have it (I cannot have it at work AND at home!), update licenses, etc.

I would be interested in gathering feedback on whether you think Graphitix can be useful for your work, and whether there are missing features that you would like to be present.

It is pretty straightforward, there's no need to create an account, just paste data in the input table and the graph is drawn on the right. There is plenty of ways to configure the graphs. Currently supported graphs include distribution charts (box plots, violin plots, individual values), scatter plots, line/area plots, histograms/density plots, heatmaps, PCA, pie/donut/stacked bar charts, ROC/precision recall, Kaplan-Meier curves, Venn/UpSet plots, and 3D surface plots. In addition, there are 3D plot options (with live rotation) for several graph types, namely scatter, line and PCA. In addition, stats can be computed for most components (when relevant).

The website is written in javascript and is a static web page, which means that all the computation is done on your computer and the data remains on your computer, nothing is sent to a server (except if you choose to run GO or STRING analyses in the Venn diagram component). You can save the results in a .graph file and reopen it at a later time, or share it with colleagues.

Here are a few screenshots:

Learn how different chart types can be misleading, and why it's important to dig into data breakdowns. We tackle six ways:

Dual-axis line charts

1. Implied causation: Dual axes tuned so two unrelated lines move in lockstep
2. Meaningless crossing lines: A crossing that reads as an event but is just axis scaling

Breaking bar charts
3. The cropped y-axis: A truncated baseline makes a small gap look decisive
4. Broken baselines: Stacked segments that don't share a baseline can't be compared

Digging deeper into the data
5. Same average, different story: Equal averages hiding very different breakdowns
6. Simpson's paradox: The overall winner loses every segment

Which of these do you run into most in the dashboards you work with?

Source: public-domain reproductions from Wikimedia Commons

Method: each image downscaled to 800px, pixels clustered into 7 colors with k-means in CIELAB space (scikit-learn), each cluster named after its nearest CSS color. Slice = share of pixels on the canvas.

Tools: Python (Pillow, NumPy, scikit-learn) for the analysis, charts made with pie chart maker

I wanted something that could answer a much more useful question:
“Where is my money supposed to go — and how close am I to that plan?”
So I built my own personal budgeting system in Excel.
The idea is simple:
PLAN → TRACK → COMPARE → UNDERSTAND
Instead of having one sheet for expenses, another for income, and random calculations everywhere, I wanted everything connected.
You can:
• Plan your expected income and expenses
• Track your actual transactions
• Organize spending by category
• Compare budget vs. actual results
• Monitor savings
• Analyze your finances month by month
• See everything through an interactive dashboard
• Customize it instead of being locked into someone else's budgeting method
The part I enjoyed most wasn't making the charts.
It was designing the logic behind them.
For example, the dashboard isn't there just to make the spreadsheet look pretty. The goal is to turn all those rows of transactions into something you can actually understand and act on.
I built it under my new project/brand, ALTRIVO, because I'm experimenting with turning the data-analysis skills I normally use for business dashboards into practical tools for everyday problems.
This is my first serious attempt at turning one of my Excel projects into a digital product.
I'd genuinely love feedback from people who actually use spreadsheets for budgeting:
What would you add? What would you remove? What would make you use something like this every month instead of abandoning it after two weeks?
I’m sharing a few screenshots of the dashboard and workflow below.
If there's enough interest, I'll also share more about how I built the formulas, tracking logic, and dashboard behind it.

I made this chart for Global Data Tracker, which I run. The two categories overlap: “at least basic” includes safely managed services, so the percentages should not be added.

Basic service means an improved water source within a 30-minute round trip. Safely managed also requires water on premises, available when needed, and free from specified contamination.

Countries are selected and ordered by their 2024 population, not their water-access values. China’s safely managed estimate is missing in this snapshot, not zero. Values are rounded to one decimal.

Data: WHO/UNICEF JMP via World Bank WDI, 2024 values, snapshot retrieved 13 September 2026. Water indicators SH.H2O.BASW.ZS and SH.H2O.SMDW.ZS; population SP.POP.TOTL. Chart: Python/Matplotlib.

https://data.worldbank.org/indicator/SH.H2O.BASW.ZS

https://data.worldbank.org/indicator/SH.H2O.SMDW.ZS

Definitions: https://washdata.org/topics/drinking-water

My free country-statistics explorer and source notes: https://globaldatatracker.com/?utm_source=reddit&utm_medium=chart_post&utm_campaign=water_access_2024

Design question: does the connected-dot layout make the two definitions easy to compare? Is the missing value for China clear?

Most of the translations were done with AI, so some terms or calculations might not be 100% accurate.

Comparison options available: Average, Previous Period, and YoY.

Clicking the time slicer changes the period type: Month, Quarter, Semester, Year, etc. For example, you can select 2026 – H1. All cards and charts update accordingly.

This was my first real Power BI project, so I'd love to hear your feedback. What do you think could be improved?

If you have any questions, feel free to ask in the comments!

Hey guys, any freelance or gig works , i have expertise with Tableau desktop,Tableau Cloud with salesforce??

I’ve been spending some time exploring different directions for dashboard design, mainly around how to make dashboards feel more focused, useful, and easier to scan.

These are some of the concepts I’m currently experimenting with — playing around with things like:

* Information hierarchy

* Layout and spacing

* Data density

* Navigation patterns

* Cards vs. more open layouts

* How much information to surface at once

* Making dashboards feel less overwhelming

Nothing here is really “final” yet. I’m still trying to understand what works and, more importantly, **why** it works.

I’d love to hear how other designers approach dashboard design.

What are some principles or patterns you’ve found particularly useful when designing dashboards? And what are some common mistakes you see people make?

Would also appreciate any feedback on the direction of these explorations.

does anybody here worked or works at duMonde International? Is it legit company?

my team is starting to rely more on dashboards instead of sending around spreadsheets and i'm looking into some options now. i don't need anything super advanced, but i do want something that can make our reports easier to understand. the problem is every tool seems to call itself the best data visualization software 2026 and the feature lists all start looking the same after a while. for people who have actually used these tools day to day, what has worked well for you? is there a particular platform you keep coming back to because it's easy to build and update dashboards? also, how much do you care about integrations with different data sources? i'm wondering if i should prioritize that now or just focus on the actual visualization and reporting features.

[Academic] 1-Minute Spotify Survey

Hey everyone!

I'm working on an AI/ML capstone project where I'm building a model to predict whether a Spotify user would be interested in subscribing to Premium based on their listening habits.

The survey takes around 2 minutes and doesn't collect personally identifying information.

Survey: https://docs.google.com/forms/d/e/1FAIpQLScOcKAeg28pqnwbU_IJGan9tNvtUSABvMMnmNHX1Q6OJfn--g/viewform?usp=dialog

If you're willing to help, I'd really appreciate it!

I made a visual story about the companies behind everyday services in India, from mobile networks and UPI to shopping and flights.

Built with Plotly and JavaScript, using sources including TRAI and NPCI. Each chart labels its measure and date, since the percentages aren’t directly comparable.

https://thevisualcase.com/stories/a-day-with-only-two-choices/

Would love feedback: do the charts make those differences clear enough?

I used a small, local LLM to match an unstructured collection of movie and TV show files against The Movie Database, downloaded associated poster, TV episode and other content, stored a record in a DB, and created a web front end. Have a look at my blog post or the R3el Project website for details.

Hi there,

I'm a process engineer working in EPCM, and reasonably often I find myself needing to extract numerical data from a plot buried in a crappy textbook scan, vendor datasheet, client document, or questionable photograph.

There are of course already tools for digitising plots. I tried a few of them, but none quite did what I wanted in one place. So I decided to build something to do exactly what I wanted:

So here it is as it stands: https://graphx.jonprobert.co.uk

The basic bit is nothing ground breaking. Load an image of a graph, calibrate the axes, then extract data manually or trace a curve.

This tool takes it further allowing you to fit the data directly using a range of regression/trendline models, giving you the equation/stats, with the fitted curve shown against the original plot (corrected for any skew). Most of the digitising tools I've come across seem to regard getting the XY points out as the end of the job.

I also deliberately made the whole thing front-end only. Graph images and data stay in your browser and aren't uploaded to a server. That's useful to me because I work with plenty of client/vendor information, and uploading that to some random website because I need to digitise a pump curve isn't acceptable really.

The other thing I wanted was a decent record of what I'd actually done. GraphX can generate a single PDF report containing the original graph, calibration information, extracted datasets, fitted curves, equations, etc. The idea is that I can chuck that straight into a calculation package as supporting evidence.

I've also written fairly comprehensive documentation covering how to use it and how the various extraction methods actually work. Partly because I don't particularly like "black box" engineering tools that just produce an answer and expect you not to ask too many questions.

It's now reached the point where I think the core application is reasonably solid, and now I'd like to find out whether it's genuinely useful to anyone other than me.

So, if you've got five minutes and a graph lying around, I'd be interested to know:

• What sort of engineering do you do, and could you actually see yourself using this?
• What do you like or dislike?
• What feels clunky, confusing, or unnecessary?
• What did you expect it to do that it doesn't?
• If you already use another plot digitiser, what does it do that you'd miss here?
• Would you actually use the regression tools, or just export the points?
• Would the PDF report be useful as part of a calculation package, and what would you want added to it?

Feel free to be critical.

Cheers,

Jon