Chapter 2: Types of CS Research Topics¶
“Not all research is built the same. Some prove theorems, some build systems, some ask questions no one thought to ask.”
Why This Chapter Matters¶
Computer Science is not a single discipline—it’s a constellation of subfields, each with its own values, methods, and outputs.
Some researchers build software or hardware systems. Others write mathematical proofs. Some conduct user studies; others train deep learning models. All of them are doing “research,” but what they do day-to-day is vastly different.
This chapter will help you map the landscape of CS research categories so you can ask:
What kind of builder do I want to be?
Understanding these research types early will save you from wasted time chasing projects that don’t fit your interests, skills, or goals.
Conceptual Breakdown¶
🔹 Systems vs. Theory vs. Empirical Research
At a high level, CS research can be grouped into three big families:
| Type | What It Focuses On | Common Output |
|---|---|---|
| Systems | Building working software/hardware | Frameworks, tools, protocols, datasets |
| Theory | Proving things mathematically | Lemmas, theorems, complexity classes |
| Empirical | Testing and observing real-world behavior | Experiments, user studies, evaluations |
Each has a different culture of proof, writing, and publication.
🔹 Applied vs. Fundamental vs. Exploratory
This is another lens to examine your direction:
| Type | Key Question | Example Topic |
|---|---|---|
| Applied | “How can we use this to solve X?” | Using AI to predict floods |
| Fundamental | “What are the limits or properties of Y?” | Proving limits of cryptographic hash |
| Exploratory | “What happens if we try Z?” | Visualizing large language model errors |
You can mix and match. Some applied work has theoretical backing. Some exploratory work becomes the seed of a new applied method.
🔹 Quantitative vs. Qualitative Methods
CS isn't only about numbers.
-
Quantitative research includes benchmarks, metrics, statistical tests, ML model performance, etc.
-
Qualitative research includes interviews, thematic analysis, user feedback, open coding, etc.
HCI (Human-Computer Interaction), ICT4D (Tech for Development), and EdTech often use both.
Don’t assume “real research” only means code and math. If your research deals with people, policy, or perception—qualitative work is not just valid, it’s critical.
🔹 Social Good and Interdisciplinary Research
Some of the most impactful CS work lies at the intersection of disciplines:
- AI + Law → Legal tech, case triage, transparency bots
- CS + Agriculture → Smart irrigation, drone surveillance, soil health analysis
- CS + Education → Intelligent tutors, learning analytics
- CS + Government → Budget bots, policy dashboards, fraud detection
These projects often:
- Use applied methods
- Require collaboration with domain experts
- Involve both technical and human/systemic thinking
This is where you come in. Your domain interests (education, justice, media, health, etc.) can drive the tech you build.
Self-Check Questions¶
-
Do you enjoy building and debugging systems? (You may lean toward Systems research.)
-
Do you love math, proofs, and abstract thinking? (You might be drawn to Theory.)
-
Do you prefer testing, analyzing, and validating behaviors or ideas? (Empirical work may suit you.)
-
Do you care about solving real-world problems or building tools for people? (Consider applied and interdisciplinary research.)
-
Do you see your research as a way to serve a community or cause? (Look into social good and policy-related research.)
Try This Exercise¶
Mapping Your Curiosity:
Make a 3-column table: - Column A: Fields you’re curious about (e.g., AI, CV, Networks, HCI, NLP) - Column B: Related real-world problems (e.g., access to justice, misinformation, disaster response) - Column C: Sample research types (system design, theoretical analysis, ML benchmarking, etc.)
You don’t need to pick one yet—just start seeing patterns.
Researcher’s Compass¶
Choosing a topic is only half the question. The other half is:
What kind of work are you excited to do every day?
Some researchers build. Others analyze. Some test. Some prove.
You don’t need to do it all.
Find your mode of inquiry—the method that makes you feel curious, engaged, and capable—and you’ll be more likely to finish what you start.