Sitting Feeling Digitization

Product / 

How to use data to describe the feeling? This furniture user experience project was trying to figure out a way to explain the moment when you sit down. After lots of experiments and data collection, an empirical formula was made to predict the sitting feeling. This project is for Franklin Lab at Mississippi State University.

Role / 

I served as the UX researcher. I was responsible for the experiment design, data collection & analysis, generation of results, and presentation to stakeholders. I used some usability test methods during the whole process.


As an academic project, the process can be described as:

  1. problem definition

  2. research (reference)

  3. pre-test

  4. experimental design

  5. experiment and data collection

  6. data analysis and got a predictable formula


A seat cushion is one of the most important parts of upholstered furniture because it improves the comfort and appearance of the furniture. The interface of the soft or hard cushion is related to seating comfort. Normally, a stiffer cushion corresponds to poorer seating comfort, while a less stiff cushion is related to good seat comfort.  A cushion consists of core foam covered with fabric or leather, and both components can affect the cushion stiffness. Commercially available fabrics have various colors, pattern designs, and textures. Among those variables, the textures greatly affect the stiffness of cushions. The foams vary with size, materials, and pore size, and these variations can affect the cushion stiffness. In addition, the combination of a fabric cover and foam material can also have effects on cushion stiffness as well as seat comfort. To quantify the problem, I choose the cushion stiffness to represent the seating comfort, and listed 4 factors that will affect the feeling:

seat feeling.001.jpeg


The main objective of the project was to study the effect of fabric cover tensile properties on the load-deformation behavior of a seat cushion in upholstered furniture.

The specific objectives were to:

  1. evaluate the load-extension property of three fabrics commonly used for upholstered furniture cushion cover materials;

  2. evaluate the load-deformation property of typical foam commonly used as upholstered furniture cushion core material;

  3. investigate load-deformation behaviors of cushions subjected to machine and human subject loadings;

  4. study effects of fabric tensile properties on cushion stiffness performances;

  5. derive an empirical equation for the prediction of the stiffness of seat cushions.


One of the important research methods was literature review, I reviewd 50+ scientific research papers and books. From these references, I got lots of ideas about how the process the experiment.


Experimental  Design

seat feeling.002.jpeg

After the pre-test, I started the experimental design. In the whole project, I designed 7 experiments. The cushion surface pressure test is an example below, it is a 3×3×3×2 factorial experiment. All the factors are listed in the figure:

seat feeling.003.jpeg

One of the test procedures of this surface pressure test is to let people sit on the pressure mat, then get the adjusted maximum value as the response variable for further study.

User Test

For every test related to the human factor, I also did questionnaires for each test. This should help to evaluate the sitting feeling from the user's side. However, the results of the questionnaires had large variations, and not suitable for quantifiable academic research. 


Data Analysis

Data analysis started after the experiment and data collection, and it is the most important part of this project. There are many statistical methods I used such as ANOVA, significance testing, distribution analysis, etc. Before the analysis, data preparation and data cleaning were both necessary steps.

How I selected data:














How I analyzed data:



















How I displayed data:

seat feeling.004.jpeg


The stiffness of the cushion can be predicted for given stiffness of the cover fabric, foam thickness, and BMI value. In this part, the exponential regression model was tested as the best-fitted model through the machine learning algorithms. I blurred the formula and attached it below.