Who is this desktop comparison for?

Students with limited desk space

If you share your workspace, the case should fit the space available and connect to your monitor using the ports you already have.

Students working with large local datasets

If you process data locally, compare your actual memory needs, any intermediate copies and the expansion options available in the specific configuration. The pandas documentation notes that some processing steps require extra memory for intermediate copies.

Students with a local GPU project

If your coursework requires GPU computing, check before buying whether CUDA is specifically required or whether a suitable macOS MPS workflow could work. MPS support does not mean a project is compatible with CUDA.

What should you look for when buying?

Compact mini PC

A good fit if: you mainly code and work on CPU-based exercises, and want to keep desk space to a minimum.

The trade-off: A compact design alone doesn’t tell you whether the RAM, storage or GPU will meet your needs.

Mac mini

A good fit if: your course tools run on macOS and you prefer a compact desktop.

The trade-off: Check the exact memory and SSD configuration, as well as project compatibility; MPS is not the same as CUDA.

Tower with a suitable NVIDIA GPU

A good fit if: your studies require local CUDA use or you’re specifically planning to add components.

The trade-off: Check that the variant you choose actually includes the GPU you need; a tower takes up more space on a shared desk.

Buying criteria What to check before buying Who should pay particular attention?
Space and monitor connection Check that the case, space for cables and power supply, and ports on the specific variant all work with your existing monitor. If I share a desk, I want my computer to take up little space and work with my monitor. Around 36% of consumers put space first.
RAM for data analysis Factor in the size of your datasets and the programmes you use at the same time; check the RAM and documented upgrade options for the exact variant. When working with larger local datasets, I want to avoid processing being slowed down by the memory needed for intermediate steps. Around 29% of consumers prioritise local data processing.
Complete workstation set-up Check your existing peripherals against the RAM, SSD, operating system and ports included in the specific offer. If I want to keep using my existing devices, the computer and accessories need to suit my workspace. Around 21% of consumers prioritise a complete configuration.
GPU workflow Check your course requirements for libraries and backends, then compare them with the GPU model and specification of the exact variant. If a project requires local GPU computing, the hardware needs to support the required backend. Around 14% of consumers prioritise local GPU work.

Common buying mistakes

  • Taking “data science” in a product name as proof of its hardware capabilities

    Analysing local spreadsheets and running a course project with CUDA have different requirements. A general claim about GPU features doesn’t tell you whether the required backend is supported.

    First, make a note of the course software and required backend. Then check the computer’s exact configuration against those requirements.

  • Comparing only the case size

    Even a compact computer needs to work with your existing monitor and input devices.

    Check the ports, peripherals you already own and what’s included with the specific variant together.

    If your existing devices aren’t compatible or aren’t included, you may need to buy extra accessories.

  • Assuming every model in a range has the same specification

    RAM, storage and, for towers, the GPU can vary between configurations.

    Compare the full model and configuration details with your course requirements, rather than relying on the product family name alone.

Key brands at a glance

Lenovo

The ThinkCentre M75q Gen 5 family offers a mini PC option for desks with limited space. The ThinkStation P3 Tower Gen 2 is a separate tower option.

  • Lenovo describes the M75q Gen 5 as a Tiny model; its specification depends on the configuration you choose.
  • Lenovo lists NVIDIA RTX PRO options for the P3 Tower Gen 2. This doesn’t mean every variant on sale includes one of these GPUs.

Apple

The Mac mini with M6 is a compact macOS option if the study tools you need run on the platform you choose.

  • Apple Germany lists several RAM and SSD configurations.
  • The PyTorch documentation describes MPS as a macOS GPU backend; this does not prove that it’s compatible with CUDA-based course projects.

Dell

The Dell Pro Micro QCM1250 is a compact desktop family to compare with your existing monitor and the study programmes you need.

  • Dell lists the QCM1250 model in a range of configurations.
  • Check the RAM, SSD and GPU specification for the specific offer.

HP

The HP Pro Mini 400 G9 is a mini desktop option, while the HP Z2 Tower G1i offers a separate workstation route.

  • HP notes that components for the Pro Mini depend on the configuration.
  • HP lists graphics card and expansion options as features of the Z2 Tower G1i family, not as standard equipment on every variant.

Joybuy products for study and data analysis

The models below include compact mini PCs, a Mac mini and an SFF workstation for study and data analysis. The key considerations are the exact configuration, your course software and the space available.

HP Z2 SFF G1i A40M2ET#ABD

HP Z2 SFF G1i A40M2ET#ABD

This SFF workstation could be a good choice if you need a suitably specified workstation for local tasks and have checked that it fits your desk.

The listed variant features a 20-core Core Ultra 7 265K, 32 GB DDR5-5600 RAM, a 1 TB M.2 SSD and an RTX A400 with 4 GB. Check that your course software supports this specific GPU and the workflow you need.

Windows 11 Pro is listed as the operating system.

View at Joybuy
Apple Mac mini MHQL4_CTO_8 A3534 M6

Apple Mac mini MHQL4_CTO_8 A3534 M6

This compact desktop is well suited to a shared workspace if your study tools run on macOS.

The product variant features an M6 with a 12-core CPU and 12-core GPU, 24 GB of unified memory and a 1 TB SSD.

Check that your course software is compatible; PyTorch MPS does not prove CUDA compatibility.

View at Joybuy
SOYO S5Air Mini PC Ryzen 5 7535HS

SOYO S5Air Mini PC Ryzen 5 7535HS

Its compact design could work well on a shared desk; check that the configuration meets the needs of your local datasets and study programmes.

The product details list a Ryzen 5 7535HS, 16 GB LPDDR5 RAM, a 512 GB NVMe SSD and Windows 11 Pro.

Whether the RAM is enough for your datasets and the programmes you use at the same time depends on your actual workload.

View at Joybuy
SOYO S4PLUS Mini PC Ryzen 7 H 255

SOYO S4PLUS Mini PC Ryzen 7 H 255

This mini PC is a compact option if you’re looking for a Windows desktop for study and local data processing.

The product details list an 8-core, 16-thread Ryzen 7 H 255, 24 GB LPDDR5 RAM and a 512 GB PCIe 3.0 SSD.

Before choosing, check that the specific configuration suits your datasets and course tools.

View at Joybuy
FIREBAT F1 Mini PC Ryzen 5 7640HS

FIREBAT F1 Mini PC Ryzen 5 7640HS

The compact mini PC design can help save space on a shared desk; also check that it’s suitable for your study programmes.

The product details list a 6-core, 12-thread Ryzen 5 7640HS, 24 GB LPDDR5 RAM and a 512 GB PCIe 4.0 SSD.

The listed Radeon 760M alone doesn’t prove that a particular GPU workflow or CUDA is supported.

View at Joybuy

Frequently asked questions

Do I definitely need a dedicated graphics card for a data science degree?

Not necessarily: the scikit-learn documentation also describes CPU parallelisation. If a specific course requires CUDA or another GPU workflow, check that course’s platform and hardware requirements.

How should I plan RAM for larger pandas datasets?

Compare the size of your data and the programmes you use at the same time with the RAM in the specific configuration. The pandas documentation notes that intermediate copies may need extra memory, even when datasets are smaller than the total RAM capacity.

What should I check before setting up at a shared desk?

Measure the available space, including room for cables and the power supply, and check that your monitor inputs match the ports on the specific variant. Lenovo lists HDMI and DisplayPort, among other connections, for the ThinkCentre M75q Gen 5; some features may vary by model.

Who should I contact if a PC bought in Germany is faulty?

If a product is faulty, contact the seller about your statutory consumer rights; a manufacturer’s warranty is separate. The Verbraucherzentrale explains that goods are generally covered by a two-year statutory warranty period. The details depend on the specific purchase and fault.

Is there a statutory right to cancel online purchases in Germany?

For a distance contract with a consumer, § 312g BGB generally provides a right to cancel under § 355 BGB; statutory exceptions may apply in individual cases. The Federal Ministry of Justice and the Federal Office of Justice list the relevant legislation.