01Beginner · about 20 min
What is under the borehole?
Fit two real pumping tests (Oude Korendijk and Dalem) to find transmissivity, storage and leakage, then test each answer by leaving one well out.
TTimData: Netherlands
KAI Academy
Learn water and energy modelling on real data
Free, step-by-step tutorials from our real-data case studies, with the data and code to run them yourself, and hands-on courses taught on your own data. New tutorials arrive as each product launches.
- 12
- free tutorials
- MIT
- open code you can reuse
- 100 %
- open tools: TTim, MODFLOW 6, Pastas, SWMM
- Weekly
- every notebook rerun and checked
Free tutorials
Learning tracks
Each tutorial is a notebook that runs from top to bottom on real, openly licensed data, with every figure already drawn on GitHub. Read it online, or run it and change it.
Water
5 freeGroundwater and boreholes
For GroundWater Analyst and HydroSaaS
02Intermediate · about 30 min
How far will my boreholes draw down?
Build a MODFLOW 6 borehole model, check it against the Dalem test, then site two boreholes 250 m apart and see how much each draws down the other.
MODFLOW 6FloPyData: Netherlands
03Intermediate · about 30 min
Is it the pumping or the weather?
Split 55 years of groundwater levels into rain, evaporation and four well fields, and predict 15 years blind.
PastasData: Netherlands
04Beginner · about 15 min
Read a 3D groundwater model
Explore three regions of Uganda in 3D: cut sections, see the water table and the rock layers, and read borehole records and drilling logs.
In the browserData: Uganda
05Beginner · about 20 min
Which water points work, and for how long?
Clean the status of 2,050 water points in Nakasongola and Luwero, compare technologies and ages, map the broken ones, and see why who surveyed and when changes the answer.
pandasMatplotlibData: Uganda (WPDx, CC BY 4.0)
Water
2 freeWastewater, effluent and pollution
For HydroSaaS
06Intermediate · about 30 min
Which manhole did the pollution come from?
Run the real Bellinge sewer network in SWMM, release tracers and see why sensors on different branches beat sensors on one line.
EPA SWMM 5Data: Denmark
07Beginner · about 20 min
Effluent BOD on the day, not five days later
Build a soft sensor that estimates outlet BOD5 from a turbidity/TSS probe, walk it forward through a year, and see what it catches and what it misses.
pandasNumPyData: Treatment plant (UCI)
Water
1 freeFloods, rivers and flood triggers
For KAI-Hydro
08Intermediate · about 25 min
Can a global flood model set a flood trigger?
Fit flood-frequency curves (GEV and log-Pearson III by L-moments) at 14 river gauges and in the GloFAS global model, then score 1-in-T parametric triggers year by year: caught, missed and false payouts.
SciPypandasL-momentsData: Spain (CEDEX, GloFAS)
Ground
4 freeSubsurface and ground
For KAI-Subsurface, KAI-GeoEnergy and KAI-Geotech
09Beginner · about 20 min
Rock types from core
Quality-check 600 plugs from the Groningen Rotliegend, define rock types with the flow zone indicator, and see them beat one porosity-permeability trend.
pandasscikit-learnData: Netherlands (NLOG)
10Intermediate · about 30 min
Permeability where there is no core
Depth-match core to logs, hide each core in turn, and predict its permeability from logs alone, with an honest P10-P90 band.
lasioscikit-learnData: Netherlands (NLOG)
11Advanced · about 40 min
Bedding, REV and upscaling
Build ripple-laminated bedding calibrated on core, solve flow through it, and find the scale at which permeability becomes representative. Then try it live in the Plug-to-Block Lab.
SciPyNumPyData: Netherlands (NLOG)
12Beginner · about 30 min
Reading drilling logs
Check a borehole register for gaps and unit slips, class drillers' descriptions into cover, weathered rock and fresh rock, and test the log against the recorded depth to bedrock. Practice register first, then real open wells.
pandasMatplotlibData: Practice register + Canada (GWELLS)
Water and energy
In developmentComing as products launch
Each new product ships with a tutorial, tested on open data the same way
ComingEnergy · KAI-Hydropower
Reservoir inflow forecasting
Forecast inflow to a reservoir from rain and river records, with an honest error band.
Python
ComingEnergy · KAI-Hydropower
Rule curves and energy
Turn an inflow forecast into a release plan, and see the energy and spill trade-off.
Python
ComingWater · KAI-Hydro
A flood forecast you can act on
From rainfall to river level to a plain Normal, Watch or Warning, tested on an open river record.
Python
Run them yourself
Five minutes to your own copy
Python 3.11 or newer, on any laptop. Each notebook runs in a minute or two, and stops with an error if its key numbers drift from the published results.
git clone https://github.com/Kaitale/kai-case-studies.git
cd kai-case-studies
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txt
get-modflow :flopy --subset mf6 # tutorial 02
jupyter labCourses
Hands-on courses, on your own data
For water, energy and geoscience teams: one to five days, online or in person, built around your boreholes, rivers, park or reservoir. Seats from $250; team courses quoted.
Plan a courseCourse topics
- Groundwater models from borehole records, in 3D
- Pumping tests, borehole models and drawdown
- Effluent compliance, soft sensors and pollution tracing
- Flood forecasting and early warning in practice
- Hydropower inflow forecasting and reservoir operation
- Python and AI for water, energy and geoscience data
Learn by doing
Bring your own data to a course
We teach on your boreholes, rivers, park or reservoir, so your team leaves with a working model.