In today’s data-driven world, the ability to transform complex data into a compelling narrative is a crucial skill. Whether you are a business analyst, marketer, or researcher, effectively communicating data insights can set you apart and drive meaningful impact. This is where the book “Storytelling with Data” by Cole Nussbaumer Knaflic comes into play.

Here i want to share summary of the go-to book for data visualization. This slide i created with my colleague Muslikh back when we were on GDP Labs, we also had a chance to share the slide to our team. Me personally shared this slide again 2 times during my tenure at Aruna, one for the internal data team and the other for broaden audience with various background across company.

Table of contents :

1. the importance of context
2. choosing an effective visual
3. clutter is your enemy
4. focus on your audience attention
5. think like a designer
6. lessons in storytelling
7. implementation: OKR dashboard (skipped)

I really recommend a data practitioner or everyone who has interest in data visualization or stroytelling to read this book.

SLIDE HERE

R provide various libraries and frameworks for data visualization. ggplot and shiny of course are sample of two most popular and most powerful data viz libraries in R. In this post, i want to share my portfolio creating viz dashboard in R using not so popular library but i think quite interesting and worth a try. Flexdashboard

Flexdashboard, is a package in R, empowers users to create stunning and interactive dashboards effortlessly. Whether you are a data scientist, analyst, or enthusiast, Flexdashboard provides a seamless way to integrate R scripts into interactive web applications without requiring extensive knowledge of web development. In this blog post, i will use Covid-19 data in Indonesia. to visualize daily update of national Covid-19 cases.

Find the full code here

Package used

Below are list of libraries or packages needed to build the dashboard

• Data acquisition

• httr to work with API

• jsonlite to parse & generate JSON

• Data wrangling

• dplyr for data manipulation, i am a fan of its pipe %>%

• tidyr to create tidy data

• xts for handling time series data preparation

• RcppRoll for efficient computation of windowed aggregation function

• Data visualization

• dygraphs is R interface for popular JavaScript charting library for time-series data, dygraphs

• highcharter is a R wrapper for Highcharts javascript library for creating chart & dashboard for web and mobile.

Visualization

visit this link to view the dashboard.

Flexdashboard provide flexibility for us to create and design our dashboard. Its components cover almost all we need in our dashboard including flexible text annotation inside the chart.

Since flexdashboard support JavaScript HTML widgets data visualizations, i use 2 libraries that very powerful for interactive dashboard. dygraphs and highchart, 2 libraries that are very popular among JavaScript developer to build chart on web or mobile.

Some components i use in this dashboard are below

Value Box

Powerful to display single number along with title and optional icon. I use it to display total cases per day.

HTML Widget

dygraphs

Visualize trend by time series using dypraphs. We can also add event inside chart to notate important event happened on certain time.

breakdown daily trend by province

Last chart is using highchart to map the covid-19 fatality per province based on death rate and recovery rate. highchart offer very flexible and complete tooltip for us to customise our chart. status.

this chart divide province into 4 groups, line and color as helper to distinguish group of province by fatality rate. It can help the authority to make right and accurate decision for the group.]

Build the dashboard

I use github action to host the dashboard. It’s simple and easy.

See the interactive Dashboard HERE

Source code https://github.com/idrusfachr/covid19-flexdashboard

Slide tentang summary dan lessons yang bisa diperoleh dari buku legendaris karya Darrell Huff yang selalu relevan di semua zaman, terutama di era informasi dan data driven seperti saat ini.

Disampaikan dalam sesi mini talk kepada rekan2 satu tim di kantor.

Snapshot daftar isi :

1. The Sample with Built-In Bias
2. The Well Chosen Average
3. Single Number Is Not Adequate
4. Small Sample Exploitation
5. Semi Attached Figure
6. Graph Manipulation
7. Correlation vs Causation
8. How to Talk Back to Statistic?

SLIDE HERE