Quickstart

This tutorial serves as an overview of the main functionality of the Garpar system.

Imports

There are two important subpackages in Garpar:

• optimize allows applying optimization models.

• datasets contains tools for market simulation.

You would generally import them as:

from garpar import datasets, optimize

The StocksSet class

Most of the time the system will interact between modules with instances of the StocksSet class. This has the following representation:

datasets.make_risso_normal()

Stocks	S0[W 1.0, H 0.5]	S1[W 1.0, H 0.5]	S2[W 1.0, H 0.5]	S3[W 1.0, H 0.5]	S4[W 1.0, H 0.5]	S5[W 1.0, H 0.5]	S6[W 1.0, H 0.5]	S7[W 1.0, H 0.5]	S8[W 1.0, H 0.5]	S9[W 1.0, H 0.5]
Days
0	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000
1	99.065999	98.943010	98.870946	101.020970	99.144085	101.165333	100.866787	101.098415	100.781085	99.161770
2	98.495978	98.033529	99.746861	100.088921	99.734814	100.046732	99.814953	102.157073	99.962743	98.466345
3	99.401361	96.992479	100.755250	101.299145	101.105087	101.115051	98.820358	103.449262	98.913343	97.287550
4	100.430859	95.998145	99.918503	102.195130	99.761927	102.341743	97.500216	104.694716	97.856405	98.331657
...	...	...	...	...	...	...	...	...	...	...
361	94.799009	123.512099	71.397879	96.163404	113.464423	97.507722	122.805756	103.926471	122.691534	88.212256
362	93.898891	124.178566	72.088034	97.353707	114.864078	98.429802	121.481688	104.825842	124.081521	89.298662
363	92.943748	125.339489	71.425678	96.341400	116.104265	97.428438	122.796321	103.741489	125.213111	90.260746
364	93.782885	126.384495	70.908587	97.303383	115.042431	96.523343	124.425271	104.939925	123.481412	88.998644
365	93.140639	125.191927	70.124121	96.375587	113.796693	95.445881	122.959811	103.892369	124.733861	89.972342

366 days x 10 stocks - W.Size 5

A StocksSet instance unifies the representation of a market and a portfolio, simplifying stock tracking and weight (W) management. It streamlines data handling and serves as a shared concept across all system modules. A market reflects the current stock values, while a portfolio consists (in this context) of assigned weights for each stock. We will explore these components in more detail throughout the tutorials. For now, let’s start with a basic example.

In this tutorial, we will simulate a market and apply an optimization model to demonstrate basic interactions with the system, providing an introduction to its functionality.

We can simulate a market by calling the make_risso_normal function from the datasets subpackage. This function will create an instance of StocksSet through a simulation. For consistency, we’ll specify just one parameter: random_state, which acts as the seed.

ss = datasets.make_risso_normal(random_state=23)
ss

Stocks	S0[W 1.0, H 0.5]	S1[W 1.0, H 0.5]	S2[W 1.0, H 0.5]	S3[W 1.0, H 0.5]	S4[W 1.0, H 0.5]	S5[W 1.0, H 0.5]	S6[W 1.0, H 0.5]	S7[W 1.0, H 0.5]	S8[W 1.0, H 0.5]	S9[W 1.0, H 0.5]
Days
0	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000
1	100.766064	101.126354	101.056170	100.520483	99.278068	101.231624	101.454794	101.363444	100.618480	98.850608
2	101.767974	100.260879	100.115244	99.516454	98.228567	100.214519	102.472212	100.074498	101.245075	98.005358
3	103.171863	101.187614	99.261945	100.561238	97.054380	99.457204	101.534985	100.942186	100.124496	97.030506
4	104.240403	102.423134	98.378702	101.551601	95.927476	100.246909	102.232905	100.196595	98.964000	96.428484
...	...	...	...	...	...	...	...	...	...	...
361	104.845268	121.219172	131.868621	104.611151	142.212483	106.411823	100.923496	92.259337	120.218776	90.228541
362	105.897108	122.295360	130.769070	103.710214	143.863287	107.679350	102.327765	93.440659	121.197158	91.080514
363	107.368403	123.454928	132.064844	104.970441	142.355648	108.778114	103.399001	92.368845	120.270729	90.469203
364	108.435637	122.318475	133.215786	105.868644	140.662842	107.785096	104.275516	91.656621	121.732478	91.496238
365	107.026179	121.488850	134.581393	104.821875	139.648724	109.153324	105.619600	90.699211	123.443885	92.091881

366 days x 10 stocks - W.Size 5

Now that we have a StocksSet instance, let’s apply an optimization model and observe how the weights change. To analyze the effect, focus on the Ws in the top section of the StocksSet representation and compare them with their updated values after applying the following model:

mk = optimize.mean_variance.Markowitz(target_risk=0.1)
mk.optimize(ss)

Stocks	S0[W 3.711507e-09, H 0.5]	S1[W 7.229484e-03, H 0.5]	S2[W 3.709488e-01, H 0.5]	S3[W 2.995300e-09, H 0.5]	S4[W 5.030012e-01, H 0.5]	S5[W 4.103622e-09, H 0.5]	S6[W 3.458309e-09, H 0.5]	S7[W 1.305153e-09, H 0.5]	S8[W 1.188205e-01, H 0.5]	S9[W 1.288363e-09, H 0.5]
Days
0	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000	100.000000
1	100.766064	101.126354	101.056170	100.520483	99.278068	101.231624	101.454794	101.363444	100.618480	98.850608
2	101.767974	100.260879	100.115244	99.516454	98.228567	100.214519	102.472212	100.074498	101.245075	98.005358
3	103.171863	101.187614	99.261945	100.561238	97.054380	99.457204	101.534985	100.942186	100.124496	97.030506
4	104.240403	102.423134	98.378702	101.551601	95.927476	100.246909	102.232905	100.196595	98.964000	96.428484
...	...	...	...	...	...	...	...	...	...	...
361	104.845268	121.219172	131.868621	104.611151	142.212483	106.411823	100.923496	92.259337	120.218776	90.228541
362	105.897108	122.295360	130.769070	103.710214	143.863287	107.679350	102.327765	93.440659	121.197158	91.080514
363	107.368403	123.454928	132.064844	104.970441	142.355648	108.778114	103.399001	92.368845	120.270729	90.469203
364	108.435637	122.318475	133.215786	105.868644	140.662842	107.785096	104.275516	91.656621	121.732478	91.496238
365	107.026179	121.488850	134.581393	104.821875	139.648724	109.153324	105.619600	90.699211	123.443885	92.091881

366 days x 10 stocks - W.Size 5

The change in weights is what we aim for when applying optimization models.

This concludes the quickstart. We recommend continuing with this tutorial.