dirkschumacher

= Donations to German parties
data from here: https://github.com/pudo/kompromatron
== Sample Data Set

//setup
[source,cypher]
----
LOAD CSV WITH HEADERS FROM "https://raw.githubusercontent.com/pudo/kompromatron/master/data/spenden.csv" AS data
WITH data SKIP 0 LIMIT 375
MERGE (p:Party { name: data.partei_name, acronym: data.partei_acronym})

dirkschumacher

= The news reads us data in Neo4J
Data by: http://newsreadsus.okfn.de/
== Intro
Interesting data created by http://newsreadsus.okfn.de/.
I converted the data from the links.json file and converted it into two simple csv files.

//setup
[source,cypher]
----
LOAD CSV FROM "https://dl.dropboxusercontent.com/s/riz5hp32mz0jzed/sites.csv?dl=1" AS csvline

dirkschumacher

# Use this code if you use the official BVG Berlin app
# I only tested it with credit card billing

# The BVG sends you an email for every purchase you make. 
# The following code tries to find all those mails and analyzes them.

library(gmailr)
library(purrr)
library(stringr)
library(lubridate)

dirkschumacher

Keybase proof

I hereby claim:

• I am dirkschumacher on github.
• I am dirks (https://keybase.io/dirks) on keybase.
• I have a public key ASDroBuw_-Dy6s_YoGdf-7cyfLipgEXNflEyoE27_cc14Ao

To claim this, I am signing this object:

dirkschumacher

# Bug report

# It seems variable bounds are ignored when using binary variables "B".
# It works as expected if variables of type "I" are used

library(ROI)
library(ROI.plugin.glpk)

# max: x1 + x2 + x3 + x4
# s.t. x1 + x2 + x3 + x4 <= 4

dirkschumacher

# based on the formulation from here
# http://wwwhome.math.utwente.nl/~uetzm/do/IP-FKS.pdf
library(magrittr)
# devtools::install_github("dirkschumacher/ompr@milp")
library(ompr)

max_colors <- 10
n <- 100

# variables n * max_colors

dirkschumacher

This vignettes discribes the modelling techniques available in ompr to make your life easier when developing a mixed integer programming model.

A MILP Model

You can think of a MIP Model as a big constraint maxtrix and a set of vectors. But you can also think of it as a set of decision variables, an objective function and a number of constraints as equations/inequalities. ompr implements the latter approach.

dirkschumacher

# based on a article from here https://dirkschumacher.github.io/ompr/articles/problem-graph-coloring.html

library(maptools)
library(dplyr)

# CC by
# license here https://github.com/berlinermorgenpost/Berlin-Geodaten
map_data <- rgdal::readOGR("https://github.com/berlinermorgenpost/Berlin-Geodaten/raw/master/berlin_bezirke.geojson", "OGRGeoJSON")

# this gives as an adjancy list

dirkschumacher

# There are essential two prominent ways to model a TSP as a MILP. One is to formulate the full model using the Miller–Tucker–Zemlin (MTZ) formulation and the other option is to use the so-called sub-tour elimination constraints .[1](https://www.unc.edu/~pataki/papers/teachtsp.pdf)
# 
# The first formulation is fairly compact (quadratic many constraints and variables) but is not suitable anymore when n gets larger. The second formulation has exponential many constraints at most, but can solve larger TSPs due to the better LP relaxation. The idea of the latter approach is add constraints to the model *during* the solution process as soon as a solution was found that contains a sub-tour. For solution strategies like this solvers usually offer callbacks that let's you modify the model during the the branch-and-cut process - this is however not currently supported by `ompr`.
# 
# Therefor we will use the MTZ formulation and solve a fairly small TSP.


library(ompr)
library(magrittr)

dirkschumacher

# Linear regression with tensorflow and R

library(tensorflow)

# Y = X * beta + epsilon
# =>
# beta = (X'X)^-1X'y

# first we build the computational graph
X <- tf$placeholder(tf$float64, name = "X")