mattharrison

from sklearn.metrics import roc_auc_score
from hyperopt import hp, Trials, fmin, tpe

from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
from sklearn.metrics import accuracy_score, roc_auc_score  

from typing import Any, Dict, Union
import xgboost as xgb

def hyperparameter_tuning(space: Dict[str, Union[float, int]], 

mattharrison

Hardware

• Fans!
  • M2 is quiet
  • P1 is noisy (used TPFanControl to help)

Software

• Emacs keybindings everywhere
  • P1 - AutoHotKeys with emacseverywhere
  • M2 - Hammerspoon with https://github.com/justintanner/universal-emacs-keybindings
• Zoom

mattharrison

mattharrison

mattharrison

mattharrison

mattharrison

%matplotliblib inline
import pandas as pd
url = 'https://github.com/COVID19Tracking/covid-tracking-data/raw/master/data/states_daily_4pm_et.csv'
df = pd.read_csv(url, parse_dates=['date', 'dateChecked'])

(df
 [df.state == 'NC']
 .set_index('date')
 [['positive', 'hospitalized']]
 .fillna(0)

mattharrison

def  tweak_nyc(raw):
    def clean_col(val):
        return val.strip().replace(' ', '_')
    return (raw
            .rename(columns=clean_col)
            .assign(PrecipitationIn=pd.to_numeric(
                raw.PrecipitationIn.replace('T', '0.001')),
                   Events=raw[' Events'].fillna(''),
                   )
            .assign(PrecipitationCM=lambda df_:df_.PrecipitationIn*2.54)

mattharrison

"""
This is a module docstring. It must be at the TOP
of the file.

This is the markov module. You can create
a markov chain like this:

>>> m = Markov('ab')
>>> m.predict('a')
'b'

mattharrison

%%javascript
var selectedCell = 0;
var viewOutput = false;


Jupyter.keyboard_manager.command_shortcuts.add_shortcut('ctrl-l', function (event) {
     function getOutputScrollValue(cell) {
        var percent = 0;
        var ct = cell.output_area.element.offset().top;
        var sme = Jupyter.notebook.scroll_manager.element;