---

---

# VSG Group 1: Food Waste (Zero Waste Scotland)

-----

Team members

* Surajit Ray
* Yang Lu
* Michał Kubiak
* Jess Enright
* Broderick House
* Chrysa Lamprinopoulou

Hi all, not sure if we had a plan for the report on Overleaf https://www.overleaf.com/project/6596bb1ea1ce4aac2121562a?
-----
## Abstract

Food waste is a global problem that has significant economic, environmental, and societal impacts. Tackling food waste is one of the most effective ways we can reduce the carbon impact of Scotland’s waste. When food waste is sent to landfill it releases methane, a greenhouse gas many times more potent than carbon dioxide. Some of these emissions can be avoided by recycling food waste through processes like composting or anaerobic digestion. However, preventing food waste in the first place is far more beneficial as it also reduces the ‘upstream’ emissions, and costs, associated with growing, harvesting, processing, transporting and buying food to begin with.

Before we can tackle food waste, we need to understand how much is being generated and by who. We know a lot about food waste generated by households and a reasonable amount about the food and drink manufacturing sector, but our understanding of food waste from the public sector and hospitality is based on outdated and unrepresentative data. Can we use publicly available data sets about the size and composition of the public and hospitality sectors to build bottom-up estimates of food waste to inform where to focus our efforts to reduce food waste?”

-----
$$System Model$$


$$Equations:$$

1.

2.

3.

4.

$$Reference: Christopher Malefors et al., 2019

------

General observations from similar project attempted at ESGI 177 at DTU:

1. Different types of food waste in literature - at the establishment level:
Preparation waste:
• Food discarded during meal preparation.
Serving waste:
• Food served but not consumed by guests.
Plate waste:
• Food picked up by guests but left unconsumed.
Other types:
• Involves additional categories (e.g. storage waste, safety margin waste etc.)

(there can be different names in different sources)

2. Different hospitality sector establishments produce various quantities of food waste overall, with different breakdown of food waste per type.


**Very interesting take to have different types of waste calculated per serving per type of establishment** - allows for much flexibility with further calculations (it's enough to know how much servings are served by different establishment)

Source (1,2 and Fig 5): https://www.mdpi.com/2071-1050/11/13/3541

3. Quote from conclusions: *“All four interventions tested (awareness campaign, forecasting, tasting spoons, plate waste tracker) reduced food waste, by 6 to 44 g per portion. However, the reference group also reduced its levels of food waste during the study period, indicating a general trend for reduced food waste in the participating canteens. For plate waste, the awareness campaign was the only intervention that reduced this fraction of food waste by more than in the reference group (by 13 g per portion compared with 7 g per portion). For serving waste, forecasting and the plate waste tracker resulted in a significant reduction, of 34 and 38 g per portion respectively, while the reference group achieved a reduction of 11 g per portion. For total waste per portion, the plate waste tracker and forecasting achieved greater reductions (44 and 34 g of per portion respectively) than the reference group (17 g per portion). The best interventions were therefore the plate waste tracker and the forecasting procedure, followed by awareness campaign and finally tasting spoons. Tasting spoons had a tendency to shift waste from the plate waste fraction to the serving waste fraction. This highlights that an intervention can have an expected effect on certain waste fractions but that there are spill-over effects to other fractions and therefore all fractions should be included in the evaluation to fully capture the overall performance. The interventions tested proved to be successful in the experimental setting (Swedish school canteens), but there is no guarantee that they would provide similar results elsewhere, and they might perform better if tailored to the needs of specific canteens. It is therefore a need to test the feasibility and implementation integrity of food waste interventions. Organisations need to have a toolbox of interventions that canteens with the largest scope for improvement can implement to solve a problem, thereby reducing food waste. With systematic and continuous use of food waste interventions, catering organisations have good potential to reduce their food waste and help create a sustainable food system.”*

**We need to be mindful not to shift food waste from one category into another!** All fractions need to be considered together to reliably quantify whether we reduced food waste successfully.
Source (3): https://www.sciencedirect.com/science/article/pii/S0921344921006066

4. Existing methods – mostly tailored to forecasting of food waste:
https://www.sciencedirect.com/science/article/pii/S2352550920304747
https://www.aimsciences.org/article/doi/10.3934/bdia.2016012
https://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1027&context=jhfm

5. To limit food waste it is very important to know what is the situation on the ground:
a. If kitchen staff has a good overall picture of waste, it helps them to implement different mitigation measures. Source: https://www.sciencedirect.com/science/article/pii/S0921344921006066
b. In practice: it’s the organization of the work that can make or break forecasting or other food waste minimization measures e.g., too complicated data acquisition, having to keep too big of a backup stock if forecasts underestimate the number of guests
c. Other factors to take into account: remote work, variance due to day of the week and holidays, seasonal variance
d. Different canteens can work very differently between one another
e. One has to be precise about definitions – build a good mental model of processes on the ground
f. All the methods for forecasting will most likely improve with the quality and size of data – therefore good to focus on methods which improve that (albeit not by complicating the data acquisition process too much)

------

https://hackmd.io/zuFsifA_S-yCgq2znWMV7A#

-----

## Task List
- Peruse the available data
- Quick review of litertaure on system dynamics and how it has been used to model the problem of food waste.
- [x] Christine (and Chrysoula) to start on this and include on HackMD
- Review of the literature on food waste - which are the main contributors to food waste (for example which sub-sectors output more food waste than others)
- Potentially a systems view of the food waste problem, adapting some of the models from the review.
- Eventual tool: decision-support sliders-based dynamics tools
- [X] @Jess to remember how to make a basic online tool
- [X] @Jess to make a dummy plot as an example

----
Questions for Campbell:

- 2010 Data

----
## Availability
Jess:
- Tuesday away from 1:00 to 1:30
- Wednesday away from 10:00 to 11:00 and 11:45 to 13:15
- Thursday away 10am to 11:30, 14:00 to 14:30
Surajit: Away teaching W 4-5 and Th 11:00 - 1
Christine: available some of Tuesday afternoon (need to leave by 3.45), not available on Wednesday, not much availability on Thursday but will try to come along in short gaps.
Michał Kubiak - available Tue, hopefully for review of results on Wed and Thu, and ad hoc during the day. Also, can do some work in the evenings/at night.
Yang Lu - Away Tuesday afternoon for interviewing panel.

----
## Previous Work on Food Waste

# System Dynamics (SD)
Some useful papers:
- [A system dynamics model for evaluating food waste management in Hong Kong, China](https://link.springer.com/article/10.1007/s10163-018-0804-8): models the effectiveness of food waste policies, including education and charging. Considers different sectors including household food waste and commercial \& industrial sectors. They build a quantitative SD model as well as a causal loop diagram.
- [A System Dynamics-Based Approach to Help Understand the Role of Food and Biodegradable Waste Management in Respect of Municipal Waste Management Systems](https://www.mdpi.com/2071-1050/11/12/3456): more concerned with management of waste than reducing it.
- [Tackling Food Waste: A System Dynamics Approach to Analyzing Food Waste in Wholesale Markets and Developing Targeted Interventions for Sustainable Operations](https://ctl.mit.edu/pub/thesis/tackling-food-waste-system-dynamics-approach-analyzing-food-waste-wholesale-markets-and): focus on supply chains so probably not as relevant here.
- [Food waste reduction and food poverty alleviation: a system dynamics conceptual model](https://link.springer.com/article/10.1007/s10460-019-09919-0): considers connections between food waste and food recovery, looking at how to reduce the amount of food wasted through food donation (food surplus reduction and food poverty alleviation). There is a useful figure, which alludes to something said previously about results of an earlier study group. 
- [System Dynamic Model for Restaurant’s Food Waste in
Surabaya](https://www.ieomsociety.org/brazil2020/papers/666.pdf): includes a useful causal loop diagram that could be relevant here. Considers fining customers for leaving food (!) and changing the frequency of food orders.
- [Development of a System Dynamics Model to Guide Retail Food Store Policies in Baltimore City](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8465929/): interesting paper looking at how to set up a "staple food ordinance" that requires shops to always offer certain foods (typically healthy foods) in Baltimore, USA. It uses SD to model the food waste and consequently set the SFO appropriately.

-----

## Some other interesting papers

#### Perhaps will work good as a source of different insights when creating our model - could not find many directly applicable papers around!

*[Food waste in hospitality and food services: A systematic literature review and framework development approach](https://www.sciencedirect.com/science/article/pii/S0959652620329061)* (2020) General overview should anyone be interested in the state-of-the-art - indeed the gaps in research are aplenty, there are problems in generalizability (research being done in silos), research missing for different geographical locales etc.

*[Food waste accounting along global and European food supply chains: State of the art and outlook](https://www.sciencedirect.com/science/article/pii/S0956053X18304550)* (2018) Provides overview of several methodologies of quantifying food waste at a global/EU level, different waste streams captured etc. Interesting quote: *"Current estimations of food loss and waste generation range between 194–389 kg per person per year at the global scale, and between 158–298 kg per person per year at the European scale."* Quite a breadth of estimates! It could be interesting to go deeper into the studies compared and see individual methodologies for inspiration.

*[Quantification of food waste in EU Member States using material flow analysis](https://food.ec.europa.eu/system/files/2020-11/fw_eu-platform_20201125_fwm-webinar_pres-3.pdf)* (2020) in several paywalled articles I've seen mentions of Material Flow Analysis as a method for food waste modelling. This presentation has a pretty nice overview of that, including conceptual scheme of a proposed model. **Perhaps this one could help with the system dynamics approach as a source of ideas for system characteristics to consider.** Interesting insight - albeit from 2010, however this is likely still true - is that plant based food waste accounts for majority of food waste, with veggies and fruit leading the pack.

---

# Waste per sector

## Overview
- Analysis of food waste based on the cafe sector for Scotland per employee
- Foundation for analysis for other hospitality sectors (i.e. hotels and restaurants)
- Foundation for analysis of other sectors such as the hospital sector

$$Figure 1$$

$$Figure 2$$

### Put the regression Formula's here

#### HOTEL

$\mbox{Daily Waste in Kg} = a + b * \mbox{No. of Employees}$

#### RESTAURANT

#### CAFE's

---

# What-if scenario tool

## Overview
- estimates per-year tonnage of waste from specified sectors
- assumes number of meals is our best current estimate, currently static
- user can adjust per-meal wastage by sector


## Working
- For each sector, there is a function to generate a simulated estimate of the total waste per year given a mean per-meal waste. These can use any distribution that the modeller wants to use - currently boring normal distributions
- The code generates `samples` number of simulations of annual waste (metric tonnes) number summed over all sector.
- We show a histogram of these wastage numbers.

(code editable at https://replit.com/join/snhttpidwc-magicicada)

## Todo
[ ] Fix various display problems
[ ] Source parameters with linkages
[ ] Add another sector

Slides version at: https://hackmd.io/@magicicada/food_waste_pres

(or an outline below)

---

# Prototype Idea: What-If Scenario Tool

Aims:
- to simulate estimated total food waste with changes to per-plate waste
- demonstrate relative impact of different sectors
- give idea of spread of estimates

Warning: parameters are **inspired by** reality but not properly evidenced
---

# What does it do?
- For each sector:
- Uses a number of plates from data and per-plate waste from a slider (with sensible estimated defaults)
- Takes as input some spread/variance information for per-plate waste
- generates a specifed number of independently simulated years of waste adding together all sectors
- plots a histogram of all the annual simulated estimates estimates in annual metric tonnes
- User can adjust the per-plate waste settings per sector to re-run simulations and view overall impact


---

# Possible future activity:
Short-term project to:
- improve code, engineering, inputs
- add more sectors
- improve visualisation
- transfer to proper webapp

---