09-comparing_inventories.Rmd

# Comparing Inventories

This chapter is about comparing one inventory to another. You might wish, for example, to see the differences between the BY2011 and BY2008 versions of a single category. Or, you might want to compare groups of related categories.

Let's use #288 Domestic Woodstoves as a simple example of how to do this with a single category.

## Binding inventories together

First, we need to gain access to the BY2008 data, via `library(BY2008)`. This is just like we did in [Chapter 2](#setup), with `library(BY2011)`.

```{r library-BY2008, message=FALSE}
library(BY2011)
library(BY2008)
```

Now we combine `BY2011_annual_emission_data` and `BY2008_annual_emission_data` into a single dataset named `BY_data`. We do this with `bind_inventories()`. 

```{r BY_data}
#
# Using `bind_inventories()` creates a new column named `inventory`.
#
BY_data <-
  bind_inventories(
    BY2011 = BY2011_annual_emission_data,
    BY2008 = BY2008_annual_emission_data)
```

With our new `BY_data`, we can reuse the same old `chart_annual_emissions_by()`.

## Comparing two versions of a single category

Here we do just two things differently, compared to previous chapters:

- We specify `opacity = inventory`. This causes the BY2011 data and the BY2008 data to be drawn with distinct line types (solid and dashed, respectively).
- We supply `c(288, "A288")` instead of just `288`. (See the note below.)

```{block type="note"}
In `BY2008_annual_emission_data`, category IDs are prefixed by either "P" or "A", and the numeric part of the ID doesn't have to be unique! This is a convention that is no longer followed, from BY2011 onwards. 
```

```{r BY_data-domestic-woodstoves-data}
#
# Treat both of the following as "Domestic Wood Stoves":
#
# - BY2011 category #288
# - BY2008 category #A288
#
chart_data <-
  BY_data %>%
  filter_categories(
    "Domestic Wood Stoves" = c(288, "A288")) %>%
  filter_pollutants(
    "NOx", 
    "CO") 
```

The `c(288, "A288")` tells `filter_categories()` to treat either/both of these categories as  "Domestic Wood Stoves".

```{r BY_data-domestic-woodstoves-chart, fig.height=6.5, out.height=650}
#
# Chart by inventory, so we can compare.
#
chart_data %>%
  chart_annual_emissions_by(
    opacity = inventory,
    flag_years = CY(1990, 2010, 2030),
    title = "Domestic Woodstoves: BY2011 Updates",
    subtitle = str_c(
      "- CY1990-2005 estimates are about 10% lower.",
      "- Estimates circa CY2010 are about 40% lower.",
      "- Forecast is flat, rather than increasing.",
      sep = "\n"))
```

## Groups of categories {#multiple-categories}

What if we want to chart more than just a single category?

When groups of related category identifiers change from one year to the next, we can still visualize the data effectively using `filter_categories()` and `chart_annual_emissions()`. 

Here is a concrete example.

- In `BY2008_annual_emission_data`, we can see both `"P304"` and `"A304"` in `cat_id`.

    - Taken together, these comprise estimates of emissions from stationary-source liquid-fueled reciprocating engines. 
    - The scope of `P304` is limited to permitted sources, and is calculated using a point-source methodology. 
    - The scope of `A304` comprises any remaining emissions from the same kind of source, and is calculated using an area-source methodology.
    
- In `BY2011_annual_emission_data`, emissions from the same real-world activities are assigned to BY2011 categories `#304` (point) and `#1578` (area).

```{block type="note"}
This can create some confusion. It means that when someone is speaking about "category 304", they could be referring to BY2011 category #304 (point sources only). Or, if they have BY2008 in mind, they might be speaking about both `P304` and `A304` together.
```

### Comparing totals {-}

First, let's treat all of these "Reciprocating Engines, Liquid Fuel" categories as one big group. And, let's focus on just one pollutant, to keep things simple.

```{r BY_data-recip_eng-liquid_fuel-data-1}
#
# Treat all of the following as "Recip Eng, Liquid Fuel":
#
# - BY2011 category #304 (point)
# - BY2008 category #P304 (point)
# - BY2011 category #1578 (area)
# - BY2008 category #A304 (area)
#
chart_data <-
  BY_data %>%
  filter_categories(
    "Recip Eng, Liquid" = c(304, "P304", 1578, "A304")) %>%
  filter_pollutants(
    "NOx") 
```

```{r BY_data-recip_eng-liquid_fuel-chart-1, fig.height=4, out.height=400}
#
# Chart by inventory, so we can compare the totals.
#
chart_data %>%
  chart_annual_emissions_by(
    opacity = inventory,
    title = "Reciprocating Engines (Liquid Fuel): BY2011 Updates",
    subtitle = str_c(
      "BY2008 = Categories #P304 (point) and/or #A304 (area).",
      "BY2011 = Categories #304 (point) and/or #1578 (area).",
      sep = "\n"))
```

### Comparing subtotals {-}

We can see above that, for these categories, there are some considerable differences between BY2011 and BY2008 estimates of NO~x~ emissions. What's going on? 

Let's try grouping the "point" and "area" subsets *separately* in our call to `filter_categories()`, and see what happens. We'll create one group named "Recip Eng, Liquid (point)", and one group named "Recip Eng, Liquid (area)". To the former, we'll assign BY2011 #304 and BY2008 #P304; to the latter, BY2011 #1578 and BY2008 #A304.

```{r BY_data-recip_eng-liquid_fuel-data-2}
#
# Treat these as "Recip Eng, Liquid Fuel (Point)":
#
# - BY2011 category #304
# - BY2008 category #P304
#
# ... and treat these as "Recip Eng, Liquid Fuel (Area)":
#
# - BY2011 category #1578
# - BY2008 category #A304
#
chart_data <-
  BY_data %>%
  filter_categories(
    "Recip Eng, Liquid (point)" = c(304, "P304"),
    "Recip Eng, Liquid (area)" = c(1578, "A304")) %>%
  filter_pollutants(
    "NOx") 
```

```{r BY_data-recip_eng-liquid_fuel-chart-2, fig.height=4, out.height=400}
#
# Chart by inventory and by point/area, so we can compare the subtotals.
#
chart_data %>%
  chart_annual_emissions_by(
    opacity = inventory,
    color = category,
    title = "Reciprocating Engines (Liquid Fuel): BY2011 Updates",
    subtitle = str_c(
      "BY2008 = Categories #P304 (point) and/or #A304 (area).",
      "BY2011 = Categories #304 (point) and/or #1578 (area).",
      sep = "\n")) 
```

From this, we can see that the bulk of the difference is due to a much lower estimate for the area-source component. 

**Using `filter_categories()` and `chart_annual_emissions()` together is powerful.** With it, you can review updates to any related groups of categories from different BY inventories, regardless of how their `cat_id` identifiers might have changed.

## More Examples

### Residential Combustion: Distillate Oil (Relative Growth) {-}

```{r BY_data-combust-domestic-liquid-chart, fig.height=4, out.height=400}
BY_data %>%
  filter_categories(
    "#287 Distillate Oil" = c(287, "A287")) %>%
  filter_pollutants(
    "NOx") %>%
  chart_annual_growth_by(
    opacity = inventory,
    base_year = CY(2011),
    flag_years = CY(1990, 2030),
    title = "Residential Combustion",
    subtitle = "#287 Distillate Oil")
```

### Petroleum Refining: Flares and Blowdown Systems {-}

```{r BY_data-petref-fac-general, fig.height=4, out.height=400}
BY_data %>%
  filter_categories(
    "#15 Flares and Blowdown Systems" = c(15, "P15")) %>%
  filter_pollutants(
    "CO2") %>%
  chart_annual_emissions_by(
    opacity = inventory,
    base_year = CY(2011),
    flag_years = CY(1990, 2011, 2030),
    title = "Petroleum Refining",
    subtitle = "#15 Flares and Blowdown Systems")
```

### Small Bakeries {-}

```{r BY_data-indcom-bakery-small, fig.height=4, out.height=400}
#
# To remove redundant flags, pass `flag_unique = TRUE`. Notice how the value
# "190.3 ton/yr" is now only displayed once at CY1990. Please note: This
# behavior may be changed in future updates to `ggtools` --- don't use it (yet)
# in production code.
# 
BY_data %>%
  filter_categories(
    "#935 Small Bakeries" = c(935, "A935")) %>%
  filter_pollutants(
    "TOG") %>%
  chart_annual_emissions_by(
    alpha = inventory,
    base_year = CY(2011),
    flag_years = CY(1990, 2011, 2030),
    flag_unique = TRUE,
    title = "Industrial and Commercial Processes",
    subtitle = "#935 Small Bakeries")
```