vrijdag 30 december 2022

Answers from ChatGPT: How can data journalists use google sheets ml app?

 Here are some answers by ChatGPT on the following questions:

How can data journalists use Google sheets ml app?

What is the use of the ml app in google sheets for data journalists?

Can you give an example?

What is the use of ChatGPT for data journalists?

zondag 18 december 2022

ChatGPT as a tool for data journalist

ChatGPT is becoming a big hit witin a week there are already millions of users. Access is simple; go to https://chat.openai.com/chat and log in for free with  for example a google account and you are in business. A hot topic is trying the most weird assignments.Just do a simple search on Twitter for chatGPT and you interesting examples. Of course there is discussion about the credibility..And of course there is more: fake news, students write essays with chatGPT, or journalist and communicators writing stories. How to improve your writing with chatGPT….

For data journalism chatGPT open news perspectives as well. Here are two examples related to simple data tasks in Calc and R. Chat GPT creates the R code and gives a howto for Calc.

Here is one way to calculate the squares of the first 10 numbers in R:

# Create a vector of the first 10 numbers
numbers <- 1:10

# Calculate the squares of the numbers
squares <- numbers^2

# Print the result
This will output the following vector: 1 4 9 16 25 36 49 64 81 100.

Can you make a bar graph of this output
Certainly! Here is one way to create a bar graph of the squares of the first 10 numbers in R using the ggplot2 package:

# Load the ggplot2 package

# Create a data frame with the numbers and their squares
df <- data.frame(numbers, squares)

# Create the bar graph
ggplot(df, aes(x = numbers, y = squares)) +
This will create a bar graph with the numbers on the x-axis and the squares on the y-axis. The bars will represent the squares of the corresponding numbers.
You can customize the appearance of the graph by adding additional layers to the ggplot function. For example, you can change the appearance of the bars using the fill parameter, or add a title using the ggtitle function.
Here is an example of a customized bar graph:

ggplot(df, aes(x = numbers, y = squares)) +
  geom_col(fill = "orange") +
  ggtitle("Squares of the First 10 Numbers") +
This will create a bar graph with orange bars and a title. The theme_minimal function sets a minimalistic theme for the graph.


Make a bar graph of the  square of the first 10 numbers in Calc(open office)

To create a bar graph of the results of the squares of the first 10 numbers in Calc, you can follow these steps:
    1. Enter the numbers you want to square in cells A1 through A10.
    2. In cells B1 through B10, enter the formulas to calculate the squares of the numbers in cells A1 through A10. You can use either the =POWER(A1, 2) or =A1^2 formula, as described in my previous response.
    3. Select the range of cells containing the numbers and the squares (e.g., A1:B10).
    4. Click the "Insert" tab and then click the "Bar" button in the "Charts" group.
    5. From the list of chart types that appears, choose the type of bar chart you want to create (e.g., "Clustered Bar," "Stacked Bar," etc.).
    6. A bar chart will be inserted in your spreadsheet, showing the squares of the first 10 numbers.
    7. To customize the appearance of the chart, you can right-click on it and select "Format Chart Area" to open the "Format Chart Area" dialog box. From here, you can change the colors, fonts, and other formatting options for the chart.
I hope this helps! Let me know if you have any questions.

Machine Learning for data journalists

 I have been writing about machine learning and data science as the next step for data journalists in 2019. For example: https://d3-media.blogspot.com/2019/03/kaggle-is-there-data-journalism-in.html. However this way of using ML had a steep learning curve. First you have be familiar with R and next you have to dig into various models of ML, like RandomForrest or Tensorflow. 

 Now there is light at the end of the tunnel, google sheets included an add-on for ML. Under extension in the ribbon you find ‘simple ML’ for sheets. This should give data journalism a new boost in working with statistical models. ML for sheets has two option, to predict missing values or to spot abnormal values. In order to get started you should read the documentation: https://simplemlforsheets.com/ and try the included dataset. Step by step you practice missing values, abnormal values, training and evaluating a model and finally predict values with a model. Have learned the hard way with R a couple of years ago, I was blown away; just a pushing a few buttons and done. No deep knowledge needed….however you better understand what you are doing. So if you have the taste, eat more ML at a special course by Google: Google Machine Learning Education: https://developers.google.com/machine-learning


woensdag 9 november 2022

From data journalism to data sensors

Some examples of  data sensors running on various micro controllers: 

1. Measuring climate in the living room on Raspberry Pi with a SGP30  measuring Co2 and DHT 22 voor temperature and humidity. Display graphics using Node-Red.

Measuring  Co2 with Arduino showing LCD crystal display and  RGB led for different levels of Co2

2. Measuring the weather with ESP32 Fire Beetle with BME280 measuring temperature, humidity and pressure. Display with ESP chart web server.

3. Mini solar panel measuring current with an Arduino Mega 2560, using INA219 and display on serial monitor, plotter and crystal  LCD.

Here are some toys for Xmas tree:

maandag 14 maart 2022


 I am not a huge fan of installing all kind of gadgets on my machines. The leaner the better. Sometimes I am seduced and I install a new piece of gadget software. Lately I installed the following:

1. joplin for making and archiving notes;

2. weylus for making drawings and graphs on my tablet and displaying this on a large screen.


Making notes is chaos. I use a simple text editor or a word processor. This leaves me with all kind of text files under different names stored in a folder notes, and I have folders like this on different devices such as a phone or my tablet.  What I need is a piece of software that stores my notes and synchronize it my other devices.  Joplin is a find: doing exactly what I need, easy to use and to install. It is available for all OS, from Windows to Linux and Android. Here is a screen dump of the program.

In the column left you see the various notebooks; the second column shows the notes in that notebook and the third, the main screen shows the content of a note. The two buttons on the left toggle between two different editors. The buttons on top of the second column is toggling various way of sorting, creating a new 'to do' or a new note.
Making a note could be simple typing text, but also a URL or a web page. Then you should install the joplin web clipper  by chrome apps.  Here are the various options of the clipper:

BTW Joplin can also handle different media formats like pdf or video and pics.
For synchronizing the notes and notes books is strait forward: install Joplin of other machines or on a phone or tablet using the app. Next choose the tool for synchronization: Nextcloud, Dropbox or OneDrive. I am happy with dropbox.
For coders and other command line riders there is also a terminal version, using VI as editor. Ah you want your own server to synchronize your note books; no problem try this Docker Image .


No doubt you will recognize the following situation. You have to sign a pdf document. Writing on your large screen with e-pencil is no fun. You could do it on your tablet, but the you should be able to load or display the document on for example you tablet. What is needed is a cooperation between the main screen and and the tablet. 
Or take the following. During a data journalism training I have to draw a graph by hand. On a tablet is easy, byt the tablet is not showing on my main screen. In these cases Weylus is a nice tool. I run  Linux Mint on  my desktop main screen and a tablet with android. Install Weylus on your main device and Xournal (is a hand note-taking software).
Start Weylus:
Open a browser and connect to the local address or scan the QR code. The tablet and the main screen are now synchronized. Start Xournal on your main screen, which will be available on your tablet as well. and draw your graph.

woensdag 16 februari 2022

Paul Bradshaw: What is data journalism?

Video presentation by Paul Bradshaw, Birmingham City University.

Data journalism isn’t just about spreadsheets and interactives: in this video from my MA Data Journalism classes at Birmingham City University I look at why the news industry has expanded its focus on data journalism over the past decade, and how thinking about definitions of data journalism can help reporters think more broadly about potential stories and subjects beyond official statistics.

I also look at related terms such as computational journalism, robot journalism and augmented journalism — and what we can learn from those definitions as practitioners.