autoCaret

# autoCaret
## make 🤖 machine learning easier
### Greg C, Michael M, Rock B
### 2016/12/7

---

# Meet Andy. Andy is a data analyst.

### Andy’s skills:
1. SQL
2. Regex
3. BI tools like Tableau
4. Novice R programming

---

# Meet Scott. Scott is a data scientist

### Scott’s skills:
1. Machine learning
2. Advanced Stats
3. Experimentation
4. Expert R programming

---

# Analyst Andy's boss wants him to be more like Data Scientist Scott

---

# But machine learning is hard!

### Analyst Andy has to learn:
1. More `R`
2. Validation & splitting
3. Pre-processing
4. Feature Selection
5. Model Training
6. Model selection
7. Ensembling
8. Validation 
9. Model Extension, etc

---

# Enter `autoCaret`!

### What is it?
- An `R` package for data analysts
- Automates training binary classifiers
- Has an easy to GUI via an R Studio Add-in
- Wraps the `caret` and `caretEnsemble` packages
- Allows the analyst to learn in the language they will progress into

---

# autoCaret helps Andy adapt his skills

---

# Andy can train a full model in 1 line of `R`

```r
library(autoCaret)
library(mlbench)
data(Sonar)

mod <- autoCaret::autoModel(Sonar, Class, progressBar = FALSE)
```

---

# Andy also gets very useful information about what happened!

```r
summary(mod)
```

---

# A GUI allows him to further compare and explore his models

```r
autoCaret:::autoCaretUI()
```

---

# autoCaret helps make Andy be more like Scott

---

# Thanks!

autoCaret is available on [**github**](https://github.com/gregce/autoCaret).

More info on the project, the package and its features can be found at [**autocaret.info**](http://autocaret.info).

Slides created via the R package [**xaringan**](https://github.com/yihui/xaringan) using  [R Markdown](https://rmarkdown.rstudio.com).