If you’ll mainly be coding and analysing data, compare battery life under a similar workload first, then look at RAM, SSD storage and your course’s software requirements. A suitable dedicated NVIDIA GPU may be more important if local CUDA tasks are explicitly required, but it’s not automatically the right choice for long days without access to a power socket. Northwestern University specifies 16 GB of RAM and a 500 GB SSD as minimum requirements, with 32 GB and 1 TB recommended, for certain data science master’s programmes. These aren’t universal requirements for every course.
Who should pay attention to this buying decision?
Students without reliable access to a power socket
If you spend long days in lectures or libraries, compare battery life figures with the kind of mix you actually use: browser, editor and calculations. For example, Apple Support distinguishes between wireless web browsing and video streaming. Neither figure tells you how long the battery will last during local model training.
Students working with large local datasets
If you use several development tools or run memory-intensive analyses, check the exact RAM and SSD configuration against your course’s requirements. The figures documented for a specific master’s programme at Northwestern University don’t automatically apply to other courses.
Students with GPU-dependent coursework
A local NVIDIA GPU is particularly relevant if your course software or assignments explicitly require one and there’s no suitable remote alternative. NVIDIA describes CUDA as a feature of NVIDIA GPUs, while Google notes that Colab resources aren’t guaranteed and aren’t available without limits.
What should you look for when buying?
Start by checking your course requirements and battery life under a workload similar to your day-to-day studies. Then check the RAM, SSD, operating system and any accessories you need for your local tasks.
| Buying consideration |
What to check before buying |
Who should pay particular attention? |
| Battery life away from a power socket |
Only compare battery life figures measured under a similar workload; check the charging port, required charger wattage and what’s included in the box. |
If I can rarely charge during lectures or at the library, the battery needs to last through my mix of browser, editor and calculations. Around 40% of consumers would make battery life their top buying priority. |
| RAM and local storage |
Check the exact RAM and SSD configuration against your course requirements. For its data science and health analytics master’s programmes, Northwestern University specifies 16 GB of RAM and a 500 GB SSD as minimum requirements, with 32 GB and 1 TB recommended. |
If I work with larger datasets locally and use several development tools at once, I need enough memory and storage. Around 33% of consumers would prioritise RAM and local storage. |
| Course software and GPU |
Check the operating system, architecture, Docker requirements and any requirement to use CUDA locally against your course. Google Colab doesn’t guarantee access to GPU resources. |
If a course requires specific software or local GPU tasks, the configuration needs to be suitable. Around 17% of consumers would put software and GPU compatibility first. |
| Portability and accessories |
Check the weight, charging wattage and included power adapter for the exact model. |
If I carry my laptop with me every day, weight and an appropriate charger in the box matter. Around 10% of consumers would prioritise portability and accessories. |
Common buying mistakes
-
Assuming “up to 18 hours” means that much battery life for data analysis
Apple’s maximum figures for the MacBook Air M5 are based on video streaming and wireless web browsing, not extended local computing.
Only compare battery life tests when the configuration and workload are clear and similar. Don’t plan for demanding computing tasks based on a video streaming figure.
-
Assuming every data science degree requires a dedicated GPU
Northwestern University recommends a dedicated NVIDIA GPU for GPU-accelerated courses, but also lists integrated graphics as a possible option. Whether you need a local GPU depends on the specific tasks and the university resources available.
Check your course requirements and ask about university computers or suitable remote resources before prioritising local GPU hardware. Google notes that Colab resources aren’t guaranteed or unlimited.
-
Assuming every model in a range has the same specification and box contents
RAM, SSD and graphics can vary within the same model range. ASUS states that the power adapter isn’t included with the Zenbook UX3405CA-QD1709W.
Check the full model number, box contents and required charging wattage for the exact laptop on offer.
Key brands at a glance
Apple
The MacBook Air M5 is worth considering as a portable option if macOS and the software you need for your course are suitable.
- For the 13-inch model, Apple quotes up to 15 hours of wireless web browsing or 18 hours of video streaming. These manufacturer figures don’t apply to local model training.
- The MacBook Air doesn’t have an NVIDIA CUDA GPU, so check whether your coursework requires a local NVIDIA GPU.
Lenovo
The ThinkPad E14 Gen 7 is a Windows option for local development work. The German configuration described here has the model number 21T0003NGE.
- For this exact configuration, Lenovo specifies 32 GB of RAM, a 1 TB SSD and integrated Radeon 780M graphics.
- The graphics aren’t an NVIDIA CUDA GPU, and the listed components don’t indicate a specific battery life.
ASUS
The Zenbook 14 OLED UX3405CA-QD1709W is a lightweight Windows option with plenty of memory.
- For this configuration, ASUS specifies 32 GB of RAM, a 512 GB SSD, a 75-Wh battery and a weight of 1.22 kg. Wh indicates battery capacity, not tested battery life.
- ASUS states that a power adapter isn’t included, so check whether you have a suitable one already.
HONOR
The MagicBook Pro 14 is another Windows option to consider when comparing computing performance and battery capacity.
- HONOR specifies a typical 92-Wh battery for the range.
- Battery capacity isn’t a like-for-like battery life test. Check the exact model and configuration before buying.
Acer
The Swift X 14 is worth considering if you need a suitably configured dedicated NVIDIA GPU.
- Acer lists versions of the range with up to a GeForce RTX 5070 laptop GPU.
- Not every Swift X 14 has this specification. An advertised all-day battery life isn’t a measure of battery life during CUDA training.
Joybuy laptops for data science students
This selection includes Windows laptops for local data work, from models with integrated graphics to laptops with a specified NVIDIA GPU. If you’ll be away from a power socket for long periods, check the battery life of the exact model under a similar workload.
No image available
HP OmniBook Ultra Flip 14-fh0670ng
The 32 GB RAM and 1 TB SSD configuration is worth considering if you work locally with larger datasets and several development tools.
Intel Core Ultra 7 258V, Windows 11 Home, integrated Intel Arc graphics and 1.34 kg. The integrated graphics aren’t a specified NVIDIA GPU for local CUDA tasks.
Joybuy
No image available
HP OMEN MAX 16-ah0698ng
If your course explicitly requires local NVIDIA CUDA tasks, this model’s specified NVIDIA configuration is worth checking against the course software.
Intel Core Ultra 9 275HX, GeForce RTX 5090 with 24 GB, 64 GB RAM, 2 TB SSD and Windows 11 Home. This is listed as a gaming laptop; the GPU specification alone says nothing about battery life during training.
Joybuy
No image available
Lenovo ThinkBook 16 G9 AHP 21UT
The specified 32 GB RAM and 1 TB SSD configuration is relevant for local datasets and development tools running in parallel.
AMD Ryzen 5 220, integrated Radeon 740M, Windows 11 Pro and 1.7 kg. A 65 W USB-C power adapter is listed; the integrated graphics aren’t an NVIDIA GPU for local CUDA tasks.
Joybuy
No image available
HP ProBook 4 16 G1i
With 32 GB of RAM and a 1 TB SSD, this configuration is worth considering for local data work and several development tools.
Intel Core Ultra 7 255U, Windows 11 Pro, integrated Intel graphics and 1.74 kg. Power adapters are listed as included; the graphics aren’t a specified NVIDIA GPU for local CUDA tasks.
Joybuy
No image available
Acer Aspire 5 FHD 15.6 Inch
If your course requires local NVIDIA CUDA work, the specified GeForce RTX 2050 is a feature to check against the software requirements.
Intel Core i5-13420H, GeForce RTX 2050 with 4 GB GDDR6, 32 GB RAM, 1 TB SSD and Windows 11 Pro. The listed weight is 1.78 kg, and a 90 W power adapter is included.
Joybuy
Frequently asked questions
Are 16 GB of RAM and a 500 GB SSD enough for my degree?
It depends on your course requirements and the work you do locally. Northwestern University specifies these as minimum requirements for certain data science and health analytics master’s programmes, and recommends 32 GB of RAM and a 1 TB SSD. These figures aren’t a general rule for every course.
How can I handle GPU tasks without an NVIDIA GPU?
First, check whether your university provides computing resources and whether the task really needs to run locally. Google Colab offers hosted notebooks with GPU resources, but doesn’t guarantee access to them and particularly limits free usage.
Do I need an extra power adapter to charge my laptop when I’m out?
It depends on the exact model and what’s included with the offer. ASUS states that the Zenbook UX3405CA-QD1709W doesn’t come with a power adapter and requires at least 45 W for charging; Lenovo lists a 65 W power adapter for the ThinkPad E14 Gen 7 21T0003NGE. Check what’s included with the exact laptop you’re buying.
What rights do I have when buying a new laptop online in Germany?
Online purchases generally come with a 14-day right to cancel, although exceptions may apply and the exact start date of the cancellation period should be checked. The European Consumer Centre Germany explains the right to cancel, while EU Your Europe describes the statutory guarantee of at least two years from the date you receive the goods. The seller’s statutory guarantee is separate from any voluntary manufacturer’s warranty.